Paper collected by Wang

Abstract:
Voxel art is a distinctive stylization widely used in games and digital media, yet automated generation from 3D meshes remains challenging due to conflicting requirements of geometric abstraction, semantic preservation, and discrete color coherence. Existing methods either over-simplify geometry or fail to achieve the pixel-precise, palette-constrained aesthetics of voxel art. We introduce Voxify3D, a differentiable two-stage framework bridging 3D mesh optimization with 2D pixel art supervision. Our core innovation lies in the synergistic integration of three components: (1) orthographic pixel art supervision that eliminates perspective distortion for precise voxel-pixel alignment; (2) patch-based CLIP alignment that preserves semantics across discretization levels; (3) palette-constrained Gumbel-Softmax quantization enabling differentiable optimization over discrete color spaces with controllable palette strategies. This integration addresses fundamental challenges: semantic preservation under extreme discretization, pixel-art aesthetics through volumetric rendering, and end-to-end discrete optimization. Experiments show superior performance (37.12 CLIP-IQA, 77.90\% user preference) across diverse characters and controllable abstraction (2-8 colors, 20x-50x resolutions). Project page: https://yichuanh.github.io/Voxify-3D/

Paperid: 2, https://arxiv.org/pdf/2512.07833.pdf GitHub

Abstract:
Humans do not just see attribute similarity -- we also see relational similarity. An apple is like a peach because both are reddish fruit, but the Earth is also like a peach: its crust, mantle, and core correspond to the peach's skin, flesh, and pit. This ability to perceive and recognize relational similarity, is arguable by cognitive scientist to be what distinguishes humans from other species. Yet, all widely used visual similarity metrics today (e.g., LPIPS, CLIP, DINO) focus solely on perceptual attribute similarity and fail to capture the rich, often surprising relational similarities that humans perceive. How can we go beyond the visible content of an image to capture its relational properties? How can we bring images with the same relational logic closer together in representation space? To answer these questions, we first formulate relational image similarity as a measurable problem: two images are relationally similar when their internal relations or functions among visual elements correspond, even if their visual attributes differ. We then curate 114k image-caption dataset in which the captions are anonymized -- describing the underlying relational logic of the scene rather than its surface content. Using this dataset, we finetune a Vision-Language model to measure the relational similarity between images. This model serves as the first step toward connecting images by their underlying relational structure rather than their visible appearance. Our study shows that while relational similarity has a lot of real-world applications, existing image similarity models fail to capture it -- revealing a critical gap in visual computing.

Paperid: 3, https://arxiv.org/pdf/2512.07831.pdf GitHub GitHub

Abstract:
Recent video generation models demonstrate impressive synthesis capabilities but remain limited by single-modality conditioning, constraining their holistic world understanding. This stems from insufficient cross-modal interaction and limited modal diversity for comprehensive world knowledge representation. To address these limitations, we introduce UnityVideo, a unified framework for world-aware video generation that jointly learns across multiple modalities (segmentation masks, human skeletons, DensePose, optical flow, and depth maps) and training paradigms. Our approach features two core components: (1) dynamic noising to unify heterogeneous training paradigms, and (2) a modality switcher with an in-context learner that enables unified processing via modular parameters and contextual learning. We contribute a large-scale unified dataset with 1.3M samples. Through joint optimization, UnityVideo accelerates convergence and significantly enhances zero-shot generalization to unseen data. We demonstrate that UnityVideo achieves superior video quality, consistency, and improved alignment with physical world constraints. Code and data can be found at: https://github.com/dvlab-research/UnityVideo

Paperid: 4, https://arxiv.org/pdf/2512.07826.pdf GitHub

Abstract:
The quality and diversity of instruction-based image editing datasets are continuously increasing, yet large-scale, high-quality datasets for instruction-based video editing remain scarce. To address this gap, we introduce OpenVE-3M, an open-source, large-scale, and high-quality dataset for instruction-based video editing. It comprises two primary categories: spatially-aligned edits (Global Style, Background Change, Local Change, Local Remove, Local Add, and Subtitles Edit) and non-spatially-aligned edits (Camera Multi-Shot Edit and Creative Edit). All edit types are generated via a meticulously designed data pipeline with rigorous quality filtering. OpenVE-3M surpasses existing open-source datasets in terms of scale, diversity of edit types, instruction length, and overall quality. Furthermore, to address the lack of a unified benchmark in the field, we construct OpenVE-Bench, containing 431 video-edit pairs that cover a diverse range of editing tasks with three key metrics highly aligned with human judgment. We present OpenVE-Edit, a 5B model trained on our dataset that demonstrates remarkable efficiency and effectiveness by setting a new state-of-the-art on OpenVE-Bench, outperforming all prior open-source models including a 14B baseline. Project page is at https://github.com/lewandofskee/OpenVE.

Paperid: 5, https://arxiv.org/pdf/2512.07810.pdf GitHub

Abstract:
Future AI systems could conceal their capabilities ('sandbagging') during evaluations, potentially misleading developers and auditors. We stress-tested sandbagging detection techniques using an auditing game. First, a red team fine-tuned five models, some of which conditionally underperformed, as a proxy for sandbagging. Second, a blue team used black-box, model-internals, or training-based approaches to identify sandbagging models. We found that the blue team could not reliably discriminate sandbaggers from benign models. Black-box approaches were defeated by effective imitation of a weaker model. Linear probes, a model-internals approach, showed more promise but their naive application was vulnerable to behaviours instilled by the red team. We also explored capability elicitation as a strategy for detecting sandbagging. Although Prompt-based elicitation was not reliable, training-based elicitation consistently elicited full performance from the sandbagging models, using only a single correct demonstration of the evaluation task. However the performance of benign models was sometimes also raised, so relying on elicitation as a detection strategy was prone to false-positives. In the short-term, we recommend developers remove potential sandbagging using on-distribution training for elicitation. In the longer-term, further research is needed to ensure the efficacy of training-based elicitation, and develop robust methods for sandbagging detection. We open source our model organisms at https://github.com/AI-Safety-Institute/sandbagging_auditing_games and select transcripts and results at https://huggingface.co/datasets/sandbagging-games/evaluation_logs . A demo illustrating the game can be played at https://sandbagging-demo.far.ai/ .

Paperid: 6, https://arxiv.org/pdf/2512.07807.pdf GitHub

Abstract:
Embedding a language field in a 3D representation enables richer semantic understanding of spatial environments by linking geometry with descriptive meaning. This allows for a more intuitive human-computer interaction, enabling querying or editing scenes using natural language, and could potentially improve tasks like scene retrieval, navigation, and multimodal reasoning. While such capabilities could be transformative, in particular for large-scale scenes, we find that recent feature distillation approaches cannot effectively learn over massive Internet data due to challenges in semantic feature misalignment and inefficiency in memory and runtime. To this end, we propose a novel approach to address these challenges. First, we introduce extremely low-dimensional semantic bottleneck features as part of the underlying 3D Gaussian representation. These are processed by rendering and passing them through a multi-resolution, feature-based, hash encoder. This significantly improves efficiency both in runtime and GPU memory. Second, we introduce an Attenuated Downsampler module and propose several regularizations addressing the semantic misalignment of ground truth 2D features. We evaluate our method on the in-the-wild HolyScenes dataset and demonstrate that it surpasses existing approaches in both performance and efficiency.

Paperid: 7, https://arxiv.org/pdf/2512.07806.pdf GitHub

Abstract:
We propose Multi-view Pyramid Transformer (MVP), a scalable multi-view transformer architecture that directly reconstructs large 3D scenes from tens to hundreds of images in a single forward pass. Drawing on the idea of ``looking broader to see the whole, looking finer to see the details," MVP is built on two core design principles: 1) a local-to-global inter-view hierarchy that gradually broadens the model's perspective from local views to groups and ultimately the full scene, and 2) a fine-to-coarse intra-view hierarchy that starts from detailed spatial representations and progressively aggregates them into compact, information-dense tokens. This dual hierarchy achieves both computational efficiency and representational richness, enabling fast reconstruction of large and complex scenes. We validate MVP on diverse datasets and show that, when coupled with 3D Gaussian Splatting as the underlying 3D representation, it achieves state-of-the-art generalizable reconstruction quality while maintaining high efficiency and scalability across a wide range of view configurations.

Paperid: 8, https://arxiv.org/pdf/2512.07805.pdf GitHub GitHub

Abstract:
We present GRAPE (Group RepresentAtional Position Encoding), a unified framework for positional encoding based on group actions. GRAPE brings together two families of mechanisms: (i) multiplicative rotations (Multiplicative GRAPE) in $\mathrm{SO}(d)$ and (ii) additive logit biases (Additive GRAPE) arising from unipotent actions in the general linear group $\mathrm{GL}$. In Multiplicative GRAPE, a position $n \in \mathbb{Z}$ (or $t \in \mathbb{R}$) acts as $\mathbf{G}(n)=\exp(n\,ω\,\mathbf{L})$ with a rank-2 skew generator $\mathbf{L} \in \mathbb{R}^{d \times d}$, yielding a relative, compositional, norm-preserving map with a closed-form matrix exponential. RoPE is recovered exactly when the $d/2$ planes are the canonical coordinate pairs with log-uniform spectrum. Learned commuting subspaces and compact non-commuting mixtures strictly extend this geometry to capture cross-subspace feature coupling at $O(d)$ and $O(r d)$ cost per head, respectively. In Additive GRAPE, additive logits arise as rank-1 (or low-rank) unipotent actions, recovering ALiBi and the Forgetting Transformer (FoX) as exact special cases while preserving an exact relative law and streaming cacheability. Altogether, GRAPE supplies a principled design space for positional geometry in long-context models, subsuming RoPE and ALiBi as special cases. Project Page: https://github.com/model-architectures/GRAPE.

Paperid: 9, https://arxiv.org/pdf/2512.07802.pdf GitHub

Abstract:
Storytelling in real-world videos often unfolds through multiple shots -- discontinuous yet semantically connected clips that together convey a coherent narrative. However, existing multi-shot video generation (MSV) methods struggle to effectively model long-range cross-shot context, as they rely on limited temporal windows or single keyframe conditioning, leading to degraded performance under complex narratives. In this work, we propose OneStory, enabling global yet compact cross-shot context modeling for consistent and scalable narrative generation. OneStory reformulates MSV as a next-shot generation task, enabling autoregressive shot synthesis while leveraging pretrained image-to-video (I2V) models for strong visual conditioning. We introduce two key modules: a Frame Selection module that constructs a semantically-relevant global memory based on informative frames from prior shots, and an Adaptive Conditioner that performs importance-guided patchification to generate compact context for direct conditioning. We further curate a high-quality multi-shot dataset with referential captions to mirror real-world storytelling patterns, and design effective training strategies under the next-shot paradigm. Finetuned from a pretrained I2V model on our curated 60K dataset, OneStory achieves state-of-the-art narrative coherence across diverse and complex scenes in both text- and image-conditioned settings, enabling controllable and immersive long-form video storytelling.

Paperid: 10, https://arxiv.org/pdf/2512.07795.pdf GitHub

Abstract:
Large language models (LLMs) are increasingly deployed in settings where reasoning, such as multi-step problem solving and chain-of-thought, is essential. Yet, current evaluation practices overwhelmingly report single-run accuracy while ignoring the intrinsic uncertainty that naturally arises from stochastic decoding. This omission creates a blind spot because practitioners cannot reliably assess whether a method's reported performance is stable, reproducible, or cost-consistent. We introduce ReasonBENCH, the first benchmark designed to quantify the underlying instability in LLM reasoning. ReasonBENCH provides (i) a modular evaluation library that standardizes reasoning frameworks, models, and tasks, (ii) a multi-run protocol that reports statistically reliable metrics for both quality and cost, and (iii) a public leaderboard to encourage variance-aware reporting. Across tasks from different domains, we find that the vast majority of reasoning strategies and models exhibit high instability. Notably, even strategies with similar average performance can display confidence intervals up to four times wider, and the top-performing methods often incur higher and less stable costs. Such instability compromises reproducibility across runs and, consequently, the reliability of reported performance. To better understand these dynamics, we further analyze the impact of prompts, model families, and scale on the trade-off between solve rate and stability. Our results highlight reproducibility as a critical dimension for reliable LLM reasoning and provide a foundation for future reasoning methods and uncertainty quantification techniques. ReasonBENCH is publicly available at https://github.com/au-clan/ReasonBench .

Paperid: 11, https://arxiv.org/pdf/2512.07778.pdf GitHub

Abstract:
Most visual generative models compress images into a latent space before applying diffusion or autoregressive modelling. Yet, existing approaches such as VAEs and foundation model aligned encoders implicitly constrain the latent space without explicitly shaping its distribution, making it unclear which types of distributions are optimal for modeling. We introduce \textbf{Distribution-Matching VAE} (\textbf{DMVAE}), which explicitly aligns the encoder's latent distribution with an arbitrary reference distribution via a distribution matching constraint. This generalizes beyond the Gaussian prior of conventional VAEs, enabling alignment with distributions derived from self-supervised features, diffusion noise, or other prior distributions. With DMVAE, we can systematically investigate which latent distributions are more conducive to modeling, and we find that SSL-derived distributions provide an excellent balance between reconstruction fidelity and modeling efficiency, reaching gFID equals 3.2 on ImageNet with only 64 training epochs. Our results suggest that choosing a suitable latent distribution structure (achieved via distribution-level alignment), rather than relying on fixed priors, is key to bridging the gap between easy-to-model latents and high-fidelity image synthesis. Code is avaliable at https://github.com/sen-ye/dmvae.

Paperid: 12, https://arxiv.org/pdf/2512.07761.pdf GitHub

Abstract:
Large language models are vulnerable to jailbreak attacks, threatening their safe deployment in real-world applications. This paper studies black-box multi-turn jailbreaks, aiming to train attacker LLMs to elicit harmful content from black-box models through a sequence of prompt-output interactions. Existing approaches typically rely on single turn optimization, which is insufficient for learning long-term attack strategies. To bridge this gap, we formulate the problem as a multi-turn reinforcement learning task, directly optimizing the harmfulness of the final-turn output as the outcome reward. To mitigate sparse supervision and promote long-term attack strategies, we propose two heuristic process rewards: (1) controlling the harmfulness of intermediate outputs to prevent triggering the black-box model's rejection mechanisms, and (2) maintaining the semantic relevance of intermediate outputs to avoid drifting into irrelevant content. Experimental results on multiple benchmarks show consistently improved attack success rates across multiple models, highlighting the effectiveness of our approach. The code is available at https://github.com/xxiqiao/RL-MTJail. Warning: This paper contains examples of harmful content.

Paperid: 13, https://arxiv.org/pdf/2512.07756.pdf GitHub

Abstract:
Clinical ultrasound acquisition is highly operator-dependent, where rapid probe motion and brightness fluctuations often lead to reconstruction errors that reduce trust and clinical utility. We present UltrasODM, a dual-stream framework that assists sonographers during acquisition through calibrated per-frame uncertainty, saliency-based diagnostics, and actionable prompts. UltrasODM integrates (i) a contrastive ranking module that groups frames by motion similarity, (ii) an optical-flow stream fused with Dual-Mamba temporal modules for robust 6-DoF pose estimation, and (iii) a Human-in-the-Loop (HITL) layer combining Bayesian uncertainty, clinician-calibrated thresholds, and saliency maps highlighting regions of low confidence. When uncertainty exceeds the threshold, the system issues unobtrusive alerts suggesting corrective actions such as re-scanning highlighted regions or slowing the sweep. Evaluated on a clinical freehand ultrasound dataset, UltrasODM reduces drift by 15.2%, distance error by 12.1%, and Hausdorff distance by 10.1% relative to UltrasOM, while producing per-frame uncertainty and saliency outputs. By emphasizing transparency and clinician feedback, UltrasODM improves reconstruction reliability and supports safer, more trustworthy clinical workflows. Our code is publicly available at https://github.com/AnandMayank/UltrasODM.

Paperid: 14, https://arxiv.org/pdf/2512.07745.pdf GitHub

Abstract:
Generative diffusion models for end-to-end autonomous driving often suffer from mode collapse, tending to generate conservative and homogeneous behaviors. While DiffusionDrive employs predefined anchors representing different driving intentions to partition the action space and generate diverse trajectories, its reliance on imitation learning lacks sufficient constraints, resulting in a dilemma between diversity and consistent high quality. In this work, we propose DiffusionDriveV2, which leverages reinforcement learning to both constrain low-quality modes and explore for superior trajectories. This significantly enhances the overall output quality while preserving the inherent multimodality of its core Gaussian Mixture Model. First, we use scale-adaptive multiplicative noise, ideal for trajectory planning, to promote broad exploration. Second, we employ intra-anchor GRPO to manage advantage estimation among samples generated from a single anchor, and inter-anchor truncated GRPO to incorporate a global perspective across different anchors, preventing improper advantage comparisons between distinct intentions (e.g., turning vs. going straight), which can lead to further mode collapse. DiffusionDriveV2 achieves 91.2 PDMS on the NAVSIM v1 dataset and 85.5 EPDMS on the NAVSIM v2 dataset in closed-loop evaluation with an aligned ResNet-34 backbone, setting a new record. Further experiments validate that our approach resolves the dilemma between diversity and consistent high quality for truncated diffusion models, achieving the best trade-off. Code and model will be available at https://github.com/hustvl/DiffusionDriveV2

Paperid: 15, https://arxiv.org/pdf/2512.07730.pdf GitHub

Abstract:
Although Multimodal Large Language Models (MLLMs) have advanced substantially, they remain vulnerable to object hallucination caused by language priors and visual information loss. To address this, we propose SAVE (Sparse Autoencoder-Driven Visual Information Enhancement), a framework that mitigates hallucination by steering the model along Sparse Autoencoder (SAE) latent features. A binary object-presence question-answering probe identifies the SAE features most indicative of the model's visual information processing, referred to as visual understanding features. Steering the model along these identified features reinforces grounded visual understanding and effectively reduces hallucination. With its simple design, SAVE outperforms state-of-the-art training-free methods on standard benchmarks, achieving a 10\%p improvement in CHAIR\_S and consistent gains on POPE and MMHal-Bench. Extensive evaluations across multiple models and layers confirm the robustness and generalizability of our approach. Further analysis reveals that steering along visual understanding features suppresses the generation of uncertain object tokens and increases attention to image tokens, mitigating hallucination. Code is released at https://github.com/wiarae/SAVE.

Paperid: 16, https://arxiv.org/pdf/2512.07720.pdf GitHub

Abstract:
Generating high-fidelity upper-body 3D avatars from one-shot input image remains a significant challenge. Current 3D avatar generation methods, which rely on large reconstruction models, are fast and capable of producing stable body structures, but they often suffer from artifacts such as blurry textures and stiff, unnatural motion. In contrast, generative video models show promising performance by synthesizing photorealistic and dynamic results, but they frequently struggle with unstable behavior, including body structural errors and identity drift. To address these limitations, we propose a novel approach that combines the strengths of both paradigms. Our framework employs a 3D reconstruction model to provide robust structural and appearance priors, which in turn guides a real-time autoregressive video diffusion model for rendering. This process enables the model to synthesize high-frequency, photorealistic details and fluid dynamics in real time, effectively reducing texture blur and motion stiffness while preventing the structural inconsistencies common in video generation methods. By uniting the geometric stability of 3D reconstruction with the generative capabilities of video models, our method produces high-fidelity digital avatars with realistic appearance and dynamic, temporally coherent motion. Experiments demonstrate that our approach significantly reduces artifacts and achieves substantial improvements in visual quality over leading methods, providing a robust and efficient solution for real-time applications such as gaming and virtual reality. Project page: https://lhyfst.github.io/visa

Paperid: 17, https://arxiv.org/pdf/2512.07703.pdf GitHub

Abstract:
Large foundation models have emerged in the last years and are pushing performance boundaries for a variety of tasks. Training or even finetuning such models demands vast datasets and computational resources, which are often scarce and costly. Adaptation methods provide a computationally efficient solution to address these limitations by allowing such models to be finetuned on small amounts of data and computing power. This is achieved by appending new trainable modules to frozen backbones with only a fraction of the trainable parameters and fitting only these modules on novel tasks. Recently, the VeRA adapter was shown to excel in parameter-efficient adaptations by utilizing a pair of frozen random low-rank matrices shared across all layers. In this paper, we propose PVeRA, a probabilistic version of the VeRA adapter, which modifies the low-rank matrices of VeRA in a probabilistic manner. This modification naturally allows handling inherent ambiguities in the input and allows for different sampling configurations during training and testing. A comprehensive evaluation was performed on the VTAB-1k benchmark and seven adapters, with PVeRA outperforming VeRA and other adapters. Our code for training models with PVeRA and benchmarking all adapters is available https://github.com/leofillioux/pvera.

Paperid: 18, https://arxiv.org/pdf/2512.07702.pdf GitHub

Abstract:
Despite substantial progress in text-to-image generation, achieving precise text-image alignment remains challenging, particularly for prompts with rich compositional structure or imaginative elements. To address this, we introduce Negative Prompting for Image Correction (NPC), an automated pipeline that improves alignment by identifying and applying negative prompts that suppress unintended content. We begin by analyzing cross-attention patterns to explain why both targeted negatives-those directly tied to the prompt's alignment error-and untargeted negatives-tokens unrelated to the prompt but present in the generated image-can enhance alignment. To discover useful negatives, NPC generates candidate prompts using a verifier-captioner-proposer framework and ranks them with a salient text-space score, enabling effective selection without requiring additional image synthesis. On GenEval++ and Imagine-Bench, NPC outperforms strong baselines, achieving 0.571 vs. 0.371 on GenEval++ and the best overall performance on Imagine-Bench. By guiding what not to generate, NPC provides a principled, fully automated route to stronger text-image alignment in diffusion models. Code is released at https://github.com/wiarae/NPC.

Paperid: 19, https://arxiv.org/pdf/2512.07698.pdf GitHub

Abstract:
Understanding articulated objects is a fundamental challenge in robotics and digital twin creation. To effectively model such objects, it is essential to recover both part segmentation and the underlying joint parameters. Despite the importance of this task, previous work has largely focused on setups like multi-view systems, object scanning, or static cameras. In this paper, we present the first data-driven approach that jointly predicts part segmentation and joint parameters from monocular video captured with a freely moving camera. Trained solely on synthetic data, our method demonstrates strong generalization to real-world objects, offering a scalable and practical solution for articulated object understanding. Our approach operates directly on casually recorded video, making it suitable for real-time applications in dynamic environments. Project webpage: https://aartykov.github.io/sim2art/

Paperid: 20, https://arxiv.org/pdf/2512.07687.pdf GitHub

Abstract:
Multimodal Large Language Models (MLLMs) have demonstrated remarkable capabilities in vision-language understanding tasks. While these models often produce linguistically coherent output, they often suffer from hallucinations, generating descriptions that are factually inconsistent with the visual content, potentially leading to adverse consequences. Therefore, the assessment of hallucinations in MLLM has become increasingly crucial in the model development process. Contemporary methodologies predominantly depend on external LLM evaluators, which are themselves susceptible to hallucinations and may present challenges in terms of domain adaptation. In this study, we propose the hypothesis that hallucination manifests as measurable irregularities within the internal layer dynamics of MLLMs, not merely due to distributional shifts but also in the context of layer-wise analysis of specific assumptions. By incorporating such modifications, \textsc{\textsc{HalluShift++}} broadens the efficacy of hallucination detection from text-based large language models (LLMs) to encompass multimodal scenarios. Our codebase is available at https://github.com/C0mRD/HalluShift_Plus.

Paperid: 21, https://arxiv.org/pdf/2512.07668.pdf GitHub

Abstract:
We address the challenge of predicting human visual attention during real-world navigation by measuring and modeling egocentric pedestrian eye gaze in an outdoor campus setting. We introduce the EgoCampus dataset, which spans 25 unique outdoor paths over 6 km across a university campus with recordings from more than 80 distinct human pedestrians, resulting in a diverse set of gaze-annotated videos. The system used for collection, Meta's Project Aria glasses, integrates eye tracking, front-facing RGB cameras, inertial sensors, and GPS to provide rich data from the human perspective. Unlike many prior egocentric datasets that focus on indoor tasks or exclude eye gaze information, our work emphasizes visual attention while subjects walk in outdoor campus paths. Using this data, we develop EgoCampusNet, a novel method to predict eye gaze of navigating pedestrians as they move through outdoor environments. Our contributions provide both a new resource for studying real-world attention and a resource for future work in gaze prediction models for navigation. Dataset and code are available upon request, and will be made publicly available at a later date at https://github.com/ComputerVisionRutgers/EgoCampus .

Paperid: 22, https://arxiv.org/pdf/2512.07667.pdf GitHub

Abstract:
Large language models sometimes assert falsehoods despite internally representing the correct answer, failures of honesty rather than accuracy, which undermines auditability and safety. Existing approaches largely optimize factual correctness or depend on retraining and brittle single-layer edits, offering limited leverage over truthful reporting. We present a training-free activation steering method that weights steering strength across network depth using a Gaussian schedule. On the MASK benchmark, which separates honesty from knowledge, we evaluate seven models spanning the LLaMA, Qwen, and Mistral families and find that Gaussian scheduling improves honesty over no-steering and single-layer baselines in six of seven models. Equal-budget ablations on LLaMA-3.1-8B-Instruct and Qwen-2.5-7B-Instruct show the Gaussian schedule outperforms random, uniform, and box-filter depth allocations, indicating that how intervention is distributed across depth materially affects outcomes beyond total strength. The method is simple, model-agnostic, requires no finetuning, and provides a low-cost control knob for eliciting truthful reporting from models' existing capabilities.

Paperid: 23, https://arxiv.org/pdf/2512.07628.pdf GitHub

Abstract:
Compositionality is critical for 3D object and scene generation, but existing part-aware 3D generation methods suffer from poor scalability due to quadratic global attention costs when increasing the number of components. In this work, we present MoCA, a compositional 3D generative model with two key designs: (1) importance-based component routing that selects top-k relevant components for sparse global attention, and (2) unimportant components compression that preserve contextual priors of unselected components while reducing computational complexity of global attention. With these designs, MoCA enables efficient, fine-grained compositional 3D asset creation with scalable number of components. Extensive experiments show MoCA outperforms baselines on both compositional object and scene generation tasks. Project page: https://lizhiqi49.github.io/MoCA

Paperid: 24, https://arxiv.org/pdf/2512.07606.pdf GitHub

Abstract:
Active learning improves annotation efficiency by selecting the most informative samples for annotation and model training. While most prior work has focused on selecting informative images for classification tasks, we investigate the more challenging setting of dense prediction, where annotations are more costly and time-intensive, especially in medical imaging. Region-level annotation has been shown to be more efficient than image-level annotation for these tasks. However, existing methods for representative annotation region selection suffer from high computational and memory costs, irrelevant region choices, and heavy reliance on uncertainty sampling. We propose decomposition sampling (DECOMP), a new active learning sampling strategy that addresses these limitations. It enhances annotation diversity by decomposing images into class-specific components using pseudo-labels and sampling regions from each class. Class-wise predictive confidence further guides the sampling process, ensuring that difficult classes receive additional annotations. Across ROI classification, 2-D segmentation, and 3-D segmentation, DECOMP consistently surpasses baseline methods by better sampling minority-class regions and boosting performance on these challenging classes. Code is in https://github.com/JingnaQiu/DECOMP.git.

Paperid: 25, https://arxiv.org/pdf/2512.07599.pdf GitHub

Abstract:
Online, real-time, and fine-grained 3D segmentation constitutes a fundamental capability for embodied intelligent agents to perceive and comprehend their operational environments. Recent advancements employ predefined object queries to aggregate semantic information from Vision Foundation Models (VFMs) outputs that are lifted into 3D point clouds, facilitating spatial information propagation through inter-query interactions. Nevertheless, perception is an inherently dynamic process, rendering temporal understanding a critical yet overlooked dimension within these prevailing query-based pipelines. Therefore, to further unlock the temporal environmental perception capabilities of embodied agents, our work reconceptualizes online 3D segmentation as an instance tracking problem (AutoSeg3D). Our core strategy involves utilizing object queries for temporal information propagation, where long-term instance association promotes the coherence of features and object identities, while short-term instance update enriches instant observations. Given that viewpoint variations in embodied robotics often lead to partial object visibility across frames, this mechanism aids the model in developing a holistic object understanding beyond incomplete instantaneous views. Furthermore, we introduce spatial consistency learning to mitigate the fragmentation problem inherent in VFMs, yielding more comprehensive instance information for enhancing the efficacy of both long-term and short-term temporal learning. The temporal information exchange and consistency learning facilitated by these sparse object queries not only enhance spatial comprehension but also circumvent the computational burden associated with dense temporal point cloud interactions. Our method establishes a new state-of-the-art, surpassing ESAM by 2.8 AP on ScanNet200 and delivering consistent gains on ScanNet, SceneNN, and 3RScan datasets.

Paperid: 26, https://arxiv.org/pdf/2512.07580.pdf GitHub

Abstract:
Vision Large Language Models (VLLMs) incur high computational costs due to their reliance on hundreds of visual tokens to represent images. While token pruning offers a promising solution for accelerating inference, this paper, however, identifies a key observation: in deeper layers (e.g., beyond the 20th), existing training-free pruning methods perform no better than random pruning. We hypothesize that this degradation is caused by "vanishing token information", where visual tokens progressively lose their salience with increasing network depth. To validate this hypothesis, we quantify a token's information content by measuring the change in the model output probabilities upon its removal. Using this proposed metric, our analysis of the information of visual tokens across layers reveals three key findings: (1) As layers deepen, the information of visual tokens gradually becomes uniform and eventually vanishes at an intermediate layer, which we term as "information horizon", beyond which the visual tokens become redundant; (2) The position of this horizon is not static; it extends deeper for visually intensive tasks, such as Optical Character Recognition (OCR), compared to more general tasks like Visual Question Answering (VQA); (3) This horizon is also strongly correlated with model capacity, as stronger VLLMs (e.g., Qwen2.5-VL) employ deeper visual tokens than weaker models (e.g., LLaVA-1.5). Based on our findings, we show that simple random pruning in deep layers efficiently balances performance and efficiency. Moreover, integrating random pruning consistently enhances existing methods. Using DivPrune with random pruning achieves state-of-the-art results, maintaining 96.9% of Qwen-2.5-VL-7B performance while pruning 50% of visual tokens. The code will be publicly available at https://github.com/YahongWang1/Information-Horizon.

Paperid: 27, https://arxiv.org/pdf/2512.07571.pdf GitHub

Abstract:
This paper presents a simple method that allows to easily enhance textual pre-trained large language models with speech information, when fine-tuned for a specific classification task. A classical issue with the fusion of many embeddings from audio with text is the large length of the audio sequence compared to the text one. Our method benefits from an existing speech tokenizer trained for Audio Speech Recognition that output long sequences of tokens from a large vocabulary, making it difficult to integrate it at low cost in a large language model. By applying a simple lasso-based feature selection on multimodal Bag-of-Words representation, we retain only the most important audio tokens for the task, and adapt the language model to them with a self-supervised language modeling objective, before fine-tuning it on the downstream task. We show this helps to improve the performances compared to an unimodal model, to a bigger SpeechLM or to integrating audio via a learned representation. We show the effectiveness of our method on two recent Argumentative Fallacy Detection and Classification tasks where the use of audio was believed counterproductive, reaching state-of-the-art results. We also provide an in-depth analysis of the method, showing that even a random audio token selection helps enhancing the unimodal model. Our code is available [online](https://github.com/salocinc/EACL26SpeechTokFallacy/).

Paperid: 28, https://arxiv.org/pdf/2512.07564.pdf GitHub

Abstract:
Vision-language models (VLMs) frequently generate hallucinated content plausible but incorrect claims about image content. We propose a training-free self-correction framework enabling VLMs to iteratively refine responses through uncertainty-guided visual re-attention. Our method combines multidimensional uncertainty quantification (token entropy, attention dispersion, semantic consistency, claim confidence) with attention-guided cropping of under-explored regions. Operating entirely with frozen, pretrained VLMs, our framework requires no gradient updates. We validate our approach on the POPE and MMHAL BENCH benchmarks using the Qwen2.5-VL-7B [23] architecture. Experimental results demonstrate that our method reduces hallucination rates by 9.8 percentage points compared to the baseline, while improving object existence accuracy by 4.7 points on adversarial splits. Furthermore, qualitative analysis confirms that uncertainty-guided re-attention successfully grounds corrections in visual evidence where standard decoding fails. We validate our approach on Qwen2.5-VL-7B [23], with plans to extend validation across diverse architectures in future versions. We release our code and methodology to facilitate future research in trustworthy multimodal systems.

Paperid: 29, https://arxiv.org/pdf/2512.07544.pdf GitHub

Abstract:
As dialogue systems become increasingly important across various domains, a key challenge in persona-based dialogue is generating engaging and context-specific interactions while ensuring the model acts with a coherent personality. However, existing persona-based dialogue datasets lack explicit relations between persona sentences and responses, which makes it difficult for models to effectively capture persona information. To address these issues, we propose MoCoRP (Modeling Consistent Relations between Persona and Response), a framework that incorporates explicit relations into language models. MoCoRP leverages an NLI expert to explicitly extract the NLI relations between persona sentences and responses, enabling the model to effectively incorporate appropriate persona information from the context into its responses. We applied this framework to pre-trained models like BART and further extended it to modern large language models (LLMs) through alignment tuning. Experimental results on the public datasets ConvAI2 and MPChat demonstrate that MoCoRP outperforms existing baselines, achieving superior persona consistency and engaging, context-aware dialogue generation. Furthermore, our model not only excels in quantitative metrics but also shows significant improvements in qualitative aspects. These results highlight the effectiveness of explicitly modeling persona-response relations in persona-based dialogue. The source codes of MoCoRP are available at https://github.com/DMCB-GIST/MoCoRP.

Paperid: 30, https://arxiv.org/pdf/2512.07542.pdf GitHub

Abstract:
Most existing latent-space models for dynamical systems require fixing the latent dimension in advance, they rely on complex loss balancing to approximate linear dynamics, and they don't regularize the latent variables. We introduce RRAEDy, a model that removes these limitations by discovering the appropriate latent dimension, while enforcing both regularized and linearized dynamics in the latent space. Built upon Rank-Reduction Autoencoders (RRAEs), RRAEDy automatically rank and prune latent variables through their singular values while learning a latent Dynamic Mode Decomposition (DMD) operator that governs their temporal progression. This structure-free yet linearly constrained formulation enables the model to learn stable and low-dimensional dynamics without auxiliary losses or manual tuning. We provide theoretical analysis demonstrating the stability of the learned operator and showcase the generality of our model by proposing an extension that handles parametric ODEs. Experiments on canonical benchmarks, including the Van der Pol oscillator, Burgers' equation, 2D Navier-Stokes, and Rotating Gaussians, show that RRAEDy achieves accurate and robust predictions. Our code is open-source and available at https://github.com/JadM133/RRAEDy. We also provide a video summarizing the main results at https://youtu.be/ox70mSSMGrM.

Paperid: 31, https://arxiv.org/pdf/2512.07539.pdf GitHub

Abstract:
Traditional Transformers face a major bottleneck in long-sequence time series forecasting due to their quadratic complexity $(\mathcal{O}(T^2))$ and their limited ability to effectively exploit frequency-domain information. Inspired by RWKV's $\mathcal{O}(T)$ linear attention and frequency-domain modeling, we propose FRWKV, a frequency-domain linear-attention framework that overcomes these limitations. Our model integrates linear attention mechanisms with frequency-domain analysis, achieving $\mathcal{O}(T)$ computational complexity in the attention path while exploiting spectral information to enhance temporal feature representations for scalable long-sequence modeling. Across eight real-world datasets, FRWKV achieves a first-place average rank. Our ablation studies confirm the critical roles of both the linear attention and frequency-encoder components. This work demonstrates the powerful synergy between linear attention and frequency analysis, establishing a new paradigm for scalable time series modeling. Code is available at this repository: https://github.com/yangqingyuan-byte/FRWKV.

Paperid: 32, https://arxiv.org/pdf/2512.07533.pdf GitHub GitHub

Abstract:
We propose VulnLLM-R, the~\emph{first specialized reasoning LLM} for vulnerability detection. Our key insight is that LLMs can reason about program states and analyze the potential vulnerabilities, rather than simple pattern matching. This can improve the model's generalizability and prevent learning shortcuts. However, SOTA reasoning LLMs are typically ultra-large, closed-source, or have limited performance in vulnerability detection. To address this, we propose a novel training recipe with specialized data selection, reasoning data generation, reasoning data filtering and correction, and testing-phase optimization. Using our proposed methodology, we train a reasoning model with seven billion parameters. Through extensive experiments on SOTA datasets across Python, C/C++, and Java, we show that VulnLLM-R has superior effectiveness and efficiency than SOTA static analysis tools and both open-source and commercial large reasoning models. We further conduct a detailed ablation study to validate the key designs in our training recipe. Finally, we construct an agent scaffold around our model and show that it outperforms CodeQL and AFL++ in real-world projects. Our agent further discovers a set of zero-day vulnerabilities in actively maintained repositories. This work represents a pioneering effort to enable real-world, project-level vulnerability detection using AI agents powered by specialized reasoning models. The code is available at~\href{https://github.com/ucsb-mlsec/VulnLLM-R}{github}.

Paperid: 33, https://arxiv.org/pdf/2512.07525.pdf GitHub

Abstract:
Rotary Position Embeddings (RoPE) have become a standard for encoding sequence order in Large Language Models (LLMs) by applying rotations to query and key vectors in the complex plane. Standard implementations, however, utilize only the real component of the complex-valued dot product for attention score calculation. This simplification discards the imaginary component, which contains valuable phase information, leading to a potential loss of relational details crucial for modeling long-context dependencies. In this paper, we propose an extension that re-incorporates this discarded imaginary component. Our method leverages the full complex-valued representation to create a dual-component attention score. We theoretically and empirically demonstrate that this approach enhances the modeling of long-context dependencies by preserving more positional information. Furthermore, evaluations on a suite of long-context language modeling benchmarks show that our method consistently improves performance over the standard RoPE, with the benefits becoming more significant as context length increases. The code is available at https://github.com/OpenMOSS/rope_pp.

Paperid: 34, https://arxiv.org/pdf/2512.07504.pdf GitHub

Abstract:
Recent text-to-image models, such as Stable Diffusion, have achieved impressive visual quality, yet they often suffer from geometric inconsistencies that undermine the structural realism of generated scenes. One prominent issue is vanishing point inconsistency, where projections of parallel lines fail to converge correctly in 2D space. This leads to structurally implausible geometry that degrades spatial realism, especially in architectural scenes. We propose ControlVP, a user-guided framework for correcting vanishing point inconsistencies in generated images. Our approach extends a pre-trained diffusion model by incorporating structural guidance derived from building contours. We also introduce geometric constraints that explicitly encourage alignment between image edges and perspective cues. Our method enhances global geometric consistency while maintaining visual fidelity comparable to the baselines. This capability is particularly valuable for applications that require accurate spatial structure, such as image-to-3D reconstruction. The dataset and source code are available at https://github.com/RyotaOkumura/ControlVP .

Paperid: 35, https://arxiv.org/pdf/2512.07487.pdf GitHub

Abstract:
A robust nonproliferation regime has contained the spread of nuclear weapons to just nine states. Yet, emerging and disruptive technologies are reshaping the landscape of nuclear risks, presenting a critical juncture for decision makers. This article lays out the contours of an overlooked but intensifying technological arms race for nuclear (in)visibility, driven by the interplay between proliferation-enabling technologies (PETs) and detection-enhancing technologies (DETs). We argue that the strategic pattern of proliferation will be increasingly shaped by the innovation pace in these domains. Artificial intelligence (AI) introduces unprecedented complexity to this equation, as its rapid scaling and knowledge substitution capabilities accelerate PET development and challenge traditional monitoring and verification methods. To analyze this dynamic, we develop a formal model centered on a Relative Advantage Index (RAI), quantifying the shifting balance between PETs and DETs. Our model explores how asymmetric technological advancement, particularly logistic AI-driven PET growth versus stepwise DET improvements, expands the band of uncertainty surrounding proliferation detectability. Through replicable scenario-based simulations, we evaluate the impact of varying PET growth rates and DET investment strategies on cumulative nuclear breakout risk. We identify a strategic fork ahead, where detection may no longer suffice without broader PET governance. Governments and international organizations should accordingly invest in policies and tools agile enough to keep pace with tomorrow's technology.

Paperid: 36, https://arxiv.org/pdf/2512.07469.pdf GitHub

Abstract:
Existing video editing methods face a critical trade-off: expert models offer precision but rely on task-specific priors like masks, hindering unification; conversely, unified temporal in-context learning models are mask-free but lack explicit spatial cues, leading to weak instruction-to-region mapping and imprecise localization. To resolve this conflict, we propose VideoCoF, a novel Chain-of-Frames approach inspired by Chain-of-Thought reasoning. VideoCoF enforces a ``see, reason, then edit" procedure by compelling the video diffusion model to first predict reasoning tokens (edit-region latents) before generating the target video tokens. This explicit reasoning step removes the need for user-provided masks while achieving precise instruction-to-region alignment and fine-grained video editing. Furthermore, we introduce a RoPE alignment strategy that leverages these reasoning tokens to ensure motion alignment and enable length extrapolation beyond the training duration. We demonstrate that with a minimal data cost of only 50k video pairs, VideoCoF achieves state-of-the-art performance on VideoCoF-Bench, validating the efficiency and effectiveness of our approach. Our code, weight, data are available at https://github.com/knightyxp/VideoCoF.

Paperid: 37, https://arxiv.org/pdf/2512.07450.pdf GitHub

Abstract:
Graph neural networks (GNNs) are increasingly used to model complex patterns in graph-structured data. However, enabling them to "forget" designated information remains challenging, especially under privacy regulations such as the GDPR. Existing unlearning methods largely optimize for efficiency and scalability, yet they offer little transparency, and the black-box nature of GNNs makes it difficult to verify whether forgetting has truly occurred. We propose an explainability-driven verifier for GNN unlearning that snapshots the model before and after deletion, using attribution shifts and localized structural changes (for example, graph edit distance) as transparent evidence. The verifier uses five explainability metrics: residual attribution, heatmap shift, explainability score deviation, graph edit distance, and a diagnostic graph rule shift. We evaluate two backbones (GCN, GAT) and four unlearning strategies (Retrain, GraphEditor, GNNDelete, IDEA) across five benchmarks (Cora, Citeseer, Pubmed, Coauthor-CS, Coauthor-Physics). Results show that Retrain and GNNDelete achieve near-complete forgetting, GraphEditor provides partial erasure, and IDEA leaves residual signals. These explanation deltas provide the primary, human-readable evidence of forgetting; we also report membership-inference ROC-AUC as a complementary, graph-wide privacy signal.

Paperid: 38, https://arxiv.org/pdf/2512.07410.pdf GitHub

Abstract:
Humanoid agents are expected to emulate the complex coordination inherent in human social behaviors. However, existing methods are largely confined to single-agent scenarios, overlooking the physically plausible interplay essential for multi-agent interactions. To bridge this gap, we propose InterAgent, the first end-to-end framework for text-driven physics-based multi-agent humanoid control. At its core, we introduce an autoregressive diffusion transformer equipped with multi-stream blocks, which decouples proprioception, exteroception, and action to mitigate cross-modal interference while enabling synergistic coordination. We further propose a novel interaction graph exteroception representation that explicitly captures fine-grained joint-to-joint spatial dependencies to facilitate network learning. Additionally, within it we devise a sparse edge-based attention mechanism that dynamically prunes redundant connections and emphasizes critical inter-agent spatial relations, thereby enhancing the robustness of interaction modeling. Extensive experiments demonstrate that InterAgent consistently outperforms multiple strong baselines, achieving state-of-the-art performance. It enables producing coherent, physically plausible, and semantically faithful multi-agent behaviors from only text prompts. Our code and data will be released to facilitate future research.

Paperid: 39, https://arxiv.org/pdf/2512.07407.pdf GitHub

Abstract:
Ensuring reliable tool use is critical for safe agentic AI systems. Language models frequently produce unreliable reasoning with plausible but incorrect solutions that are difficult to verify. To address this, we investigate fine-tuning models to use Prolog as an external tool for verifiable computation. Using Group Relative Policy Optimization (GRPO), we fine-tune Qwen2.5-3B-Instruct on a cleaned GSM8K-Prolog-Prover dataset while varying (i) prompt structure, (ii) reward composition (execution, syntax, semantics, structure), and (iii) inference protocol: single-shot, best-of-N, and two agentic modes where Prolog is invoked internally or independently. Our reinforcement learning approach outperforms supervised fine-tuning, with our 3B model achieving zero-shot MMLU performance comparable to 7B few-shot results. Our findings reveal that: 1) joint tuning of prompt, reward, and inference shapes program syntax and logic; 2) best-of-N with external Prolog verification maximizes accuracy on GSM8K; 3) agentic inference with internal repair yields superior zero-shot generalization on MMLU-Stem and MMLU-Pro. These results demonstrate that grounding model reasoning in formal verification systems substantially improves reliability and auditability for safety-critical applications. The source code for reproducing our experiments is available under https://github.com/niklasmellgren/grpo-prolog-inference

Paperid: 40, https://arxiv.org/pdf/2512.07394.pdf GitHub

Abstract:
We introduce the task of Reconstructing Objects along Hand Interaction Timelines (ROHIT). We first define the Hand Interaction Timeline (HIT) from a rigid object's perspective. In a HIT, an object is first static relative to the scene, then is held in hand following contact, where its pose changes. This is usually followed by a firm grip during use, before it is released to be static again w.r.t. to the scene. We model these pose constraints over the HIT, and propose to propagate the object's pose along the HIT enabling superior reconstruction using our proposed Constrained Optimisation and Propagation (COP) framework. Importantly, we focus on timelines with stable grasps - i.e. where the hand is stably holding an object, effectively maintaining constant contact during use. This allows us to efficiently annotate, study, and evaluate object reconstruction in videos without 3D ground truth. We evaluate our proposed task, ROHIT, over two egocentric datasets, HOT3D and in-the-wild EPIC-Kitchens. In HOT3D, we curate 1.2K clips of stable grasps. In EPIC-Kitchens, we annotate 2.4K clips of stable grasps including 390 object instances across 9 categories from videos of daily interactions in 141 environments. Without 3D ground truth, we utilise 2D projection error to assess the reconstruction. Quantitatively, COP improves stable grasp reconstruction by 6.2-11.3% and HIT reconstruction by up to 24.5% with constrained pose propagation.

Paperid: 41, https://arxiv.org/pdf/2512.07391.pdf GitHub

Abstract:
Convolutional Neural Networks (CNNs) have proven highly effective for edge and mobile vision tasks due to their computational efficiency. While many recent works seek to enhance CNNs with global contextual understanding via self-attention-based Vision Transformers, these approaches often introduce significant computational overhead. In this work, we demonstrate that it is possible to retain strong global perception without relying on computationally expensive components. We present GlimmerNet, an ultra-lightweight convolutional network built on the principle of separating receptive field diversity from feature recombination. GlimmerNet introduces Grouped Dilated Depthwise Convolutions(GDBlocks), which partition channels into groups with distinct dilation rates, enabling multi-scale feature extraction at no additional parameter cost. To fuse these features efficiently, we design a novel Aggregator module that recombines cross-group representations using grouped pointwise convolution, significantly lowering parameter overhead. With just 31K parameters and 29% fewer FLOPs than the most recent baseline, GlimmerNet achieves a new state-of-the-art weighted F1-score of 0.966 on the UAV-focused AIDERv2 dataset. These results establish a new accuracy-efficiency trade-off frontier for real-time emergency monitoring on resource-constrained UAV platforms. Our implementation is publicly available at https://github.com/djordjened92/gdd-cnn.

Paperid: 42, https://arxiv.org/pdf/2512.07385.pdf GitHub GitHub

Abstract:
Unmanned Aerial Vehicles (UAVs) offer wide-ranging applications but also pose significant safety and privacy violation risks in areas like airport and infrastructure inspection, spurring the rapid development of Anti-UAV technologies in recent years. However, current Anti-UAV research primarily focuses on RGB, infrared (IR), or RGB-IR videos captured by fixed ground cameras, with little attention to tracking target UAVs from another moving UAV platform. To fill this gap, we propose a new multi-modal visual tracking task termed UAV-Anti-UAV, which involves a pursuer UAV tracking a target adversarial UAV in the video stream. Compared to existing Anti-UAV tasks, UAV-Anti-UAV is more challenging due to severe dual-dynamic disturbances caused by the rapid motion of both the capturing platform and the target. To advance research in this domain, we construct a million-scale dataset consisting of 1,810 videos, each manually annotated with bounding boxes, a language prompt, and 15 tracking attributes. Furthermore, we propose MambaSTS, a Mamba-based baseline method for UAV-Anti-UAV tracking, which enables integrated spatial-temporal-semantic learning. Specifically, we employ Mamba and Transformer models to learn global semantic and spatial features, respectively, and leverage the state space model's strength in long-sequence modeling to establish video-level long-term context via a temporal token propagation mechanism. We conduct experiments on the UAV-Anti-UAV dataset to validate the effectiveness of our method. A thorough experimental evaluation of 50 modern deep tracking algorithms demonstrates that there is still significant room for improvement in the UAV-Anti-UAV domain. The dataset and codes will be available at {\color{magenta}https://github.com/983632847/Awesome-Multimodal-Object-Tracking}.

Paperid: 43, https://arxiv.org/pdf/2512.07371.pdf GitHub

Abstract:
Behavior-cloning based visuomotor policies enable precise manipulation but often inherit the slow, cautious tempo of human demonstrations, limiting practical deployment. However, prior studies on acceleration methods mainly rely on statistical or heuristic cues that ignore task semantics and can fail across diverse manipulation settings. We present ESPADA, a semantic and spatially aware framework that segments demonstrations using a VLM-LLM pipeline with 3D gripper-object relations, enabling aggressive downsampling only in non-critical segments while preserving precision-critical phases, without requiring extra data or architectural modifications, or any form of retraining. To scale from a single annotated episode to the full dataset, ESPADA propagates segment labels via Dynamic Time Warping (DTW) on dynamics-only features. Across both simulation and real-world experiments with ACT and DP baselines, ESPADA achieves approximately a 2x speed-up while maintaining success rates, narrowing the gap between human demonstrations and efficient robot control.

Paperid: 44, https://arxiv.org/pdf/2512.07352.pdf GitHub GitHub

Abstract:
Existing speech anti-spoofing benchmarks rely on a narrow set of public models, creating a substantial gap from real-world scenarios in which commercial systems employ diverse, often proprietary APIs. To address this issue, we introduce MultiAPI Spoof, a multi-API audio anti-spoofing dataset comprising about 230 hours of synthetic speech generated by 30 distinct APIs, including commercial services, open-source models, and online platforms. Based on this dataset, we define the API tracing task, enabling fine-grained attribution of spoofed audio to its generation source. We further propose Nes2Net-LA, a local-attention enhanced variant of Nes2Net that improves local context modeling and fine-grained spoofing feature extraction. Experiments show that Nes2Net-LA achieves state-of-the-art performance and offers superior robustness, particularly under diverse and unseen spoofing conditions. Code \footnote{https://github.com/XuepingZhang/MultiAPI-Spoof} and dataset \footnote{https://xuepingzhang.github.io/MultiAPI-Spoof-Dataset/} have released.

Paperid: 45, https://arxiv.org/pdf/2512.07342.pdf GitHub

Abstract:
Recently, offline reinforcement learning (RL) has become a popular RL paradigm. In offline RL, data providers share pre-collected datasets -- either as individual transitions or sequences of transitions forming trajectories -- to enable the training of RL models (also called agents) without direct interaction with the environments. Offline RL saves interactions with environments compared to traditional RL, and has been effective in critical areas, such as navigation tasks. Meanwhile, concerns about privacy leakage from offline RL datasets have emerged. To safeguard private information in offline RL datasets, we propose the first differential privacy (DP) offline dataset synthesis method, PrivORL, which leverages a diffusion model and diffusion transformer to synthesize transitions and trajectories, respectively, under DP. The synthetic dataset can then be securely released for downstream analysis and research. PrivORL adopts the popular approach of pre-training a synthesizer on public datasets, and then fine-tuning on sensitive datasets using DP Stochastic Gradient Descent (DP-SGD). Additionally, PrivORL introduces curiosity-driven pre-training, which uses feedback from the curiosity module to diversify the synthetic dataset and thus can generate diverse synthetic transitions and trajectories that closely resemble the sensitive dataset. Extensive experiments on five sensitive offline RL datasets show that our method achieves better utility and fidelity in both DP transition and trajectory synthesis compared to baselines. The replication package is available at the GitHub repository.

Paperid: 46, https://arxiv.org/pdf/2512.07338.pdf GitHub

Abstract:
Referring expression segmentation is a fundamental task in computer vision that integrates natural language understanding with precise visual localization of target regions. Considering aerial imagery (e.g., modern aerial photos collected through drones, historical photos from aerial archives, high-resolution satellite imagery, etc.) presents unique challenges because spatial resolution varies widely across datasets, the use of color is not consistent, targets often shrink to only a few pixels, and scenes contain very high object densities and objects with partial occlusions. This work presents Aerial-D, a new large-scale referring expression segmentation dataset for aerial imagery, comprising 37,288 images with 1,522,523 referring expressions that cover 259,709 annotated targets, spanning across individual object instances, groups of instances, and semantic regions covering 21 distinct classes that range from vehicles and infrastructure to land coverage types. The dataset was constructed through a fully automatic pipeline that combines systematic rule-based expression generation with a Large Language Model (LLM) enhancement procedure that enriched both the linguistic variety and the focus on visual details within the referring expressions. Filters were additionally used to simulate historic imaging conditions for each scene. We adopted the RSRefSeg architecture, and trained models on Aerial-D together with prior aerial datasets, yielding unified instance and semantic segmentation from text for both modern and historical images. Results show that the combined training achieves competitive performance on contemporary benchmarks, while maintaining strong accuracy under monochrome, sepia, and grainy degradations that appear in archival aerial photography. The dataset, trained models, and complete software pipeline are publicly available at https://luispl77.github.io/aerial-d .

Paperid: 47, https://arxiv.org/pdf/2512.07328.pdf GitHub GitHub

Abstract:
Text-to-video (T2V) generation has advanced rapidly, yet maintaining consistent character identities across scenes remains a major challenge. Existing personalization methods often focus on facial identity but fail to preserve broader contextual cues such as hairstyle, outfit, and body shape, which are critical for visual coherence. We propose \textbf{ContextAnyone}, a context-aware diffusion framework that achieves character-consistent video generation from text and a single reference image. Our method jointly reconstructs the reference image and generates new video frames, enabling the model to fully perceive and utilize reference information. Reference information is effectively integrated into a DiT-based diffusion backbone through a novel Emphasize-Attention module that selectively reinforces reference-aware features and prevents identity drift across frames. A dual-guidance loss combines diffusion and reference reconstruction objectives to enhance appearance fidelity, while the proposed Gap-RoPE positional embedding separates reference and video tokens to stabilize temporal modeling. Experiments demonstrate that ContextAnyone outperforms existing reference-to-video methods in identity consistency and visual quality, generating coherent and context-preserving character videos across diverse motions and scenes. Project page: \href{https://github.com/ziyang1106/ContextAnyone}{https://github.com/ziyang1106/ContextAnyone}.

Paperid: 48, https://arxiv.org/pdf/2512.07314.pdf GitHub

Abstract:
Modeling human mobility is vital for extensive applications such as transportation planning and epidemic modeling. With the rise of the Artificial Intelligence Generated Content (AIGC) paradigm, recent works explore synthetic trajectory generation using autoregressive and diffusion models. While these methods show promise for generating single-day trajectories, they remain limited by inefficiencies in long-term generation (e.g., weekly trajectories) and a lack of explicit spatiotemporal multi-scale modeling. This study proposes Multi-Scale Spatio-Temporal AutoRegression (M-STAR), a new framework that generates long-term trajectories through a coarse-to-fine spatiotemporal prediction process. M-STAR combines a Multi-scale Spatiotemporal Tokenizer that encodes hierarchical mobility patterns with a Transformer-based decoder for next-scale autoregressive prediction. Experiments on two real-world datasets show that M-STAR outperforms existing methods in fidelity and significantly improves generation speed. The data and codes are available at https://github.com/YuxiaoLuo0013/M-STAR.

Paperid: 49, https://arxiv.org/pdf/2512.07302.pdf GitHub

Abstract:
Existing image perception methods based on VLMs generally follow a paradigm wherein models extract and analyze image content based on user-provided textual task prompts. However, such methods face limitations when applied to UAV imagery, which presents challenges like target confusion, scale variations, and complex backgrounds. These challenges arise because VLMs' understanding of image content depends on the semantic alignment between visual and textual tokens. When the task prompt is simplistic and the image content is complex, achieving effective alignment becomes difficult, limiting the model's ability to focus on task-relevant information. To address this issue, we introduce AerialVP, the first agent framework for task prompt enhancement in UAV image perception. AerialVP proactively extracts multi-dimensional auxiliary information from UAV images to enhance task prompts, overcoming the limitations of traditional VLM-based approaches. Specifically, the enhancement process includes three stages: (1) analyzing the task prompt to identify the task type and enhancement needs, (2) selecting appropriate tools from the tool repository, and (3) generating enhanced task prompts based on the analysis and selected tools. To evaluate AerialVP, we introduce AerialSense, a comprehensive benchmark for UAV image perception that includes Aerial Visual Reasoning, Aerial Visual Question Answering, and Aerial Visual Grounding tasks. AerialSense provides a standardized basis for evaluating model generalization and performance across diverse resolutions, lighting conditions, and both urban and natural scenes. Experimental results demonstrate that AerialVP significantly enhances task prompt guidance, leading to stable and substantial performance improvements in both open-source and proprietary VLMs. Our work will be available at https://github.com/lostwolves/AerialVP.

Paperid: 50, https://arxiv.org/pdf/2512.07276.pdf GitHub

Abstract:
Three-dimensional geospatial analysis is critical to applications in urban planning, climate adaptation, and environmental assessment. Current methodologies depend on costly, specialized sensors (e.g., LiDAR and multispectral), which restrict global accessibility. Existing sensor-based and rule-driven methods further struggle with tasks requiring the integration of multiple 3D cues, handling diverse queries, and providing interpretable reasoning. We hereby present Geo3DVQA, a comprehensive benchmark for evaluating vision-language models (VLMs) in height-aware, 3D geospatial reasoning using RGB-only remote sensing imagery. Unlike conventional sensor-based frameworks, Geo3DVQA emphasizes realistic scenarios that integrate elevation, sky view factors, and land cover patterns. The benchmark encompasses 110k curated question-answer pairs spanning 16 task categories across three complexity levels: single-feature inference, multi-feature reasoning, and application-level spatial analysis. The evaluation of ten state-of-the-art VLMs highlights the difficulty of RGB-to-3D reasoning. GPT-4o and Gemini-2.5-Flash achieved only 28.6% and 33.0% accuracy respectively, while domain-specific fine-tuning of Qwen2.5-VL-7B achieved 49.6% (+24.8 points). These results reveal both the limitations of current VLMs and the effectiveness of domain adaptation. Geo3DVQA introduces new challenge frontiers for scalable, accessible, and holistic 3D geospatial analysis. The dataset and code will be released upon publication at https://github.com/mm1129/Geo3DVQA.

Paperid: 51, https://arxiv.org/pdf/2512.07261.pdf GitHub

Abstract:
Discrete Controller Synthesis (DCS) is a powerful formal method for automatically generating specifications of discrete event systems. However, its practical adoption is often hindered by the highly specialized nature of formal models written in languages such as FSP and FLTL. In practice, syntax errors in modeling frequently become an important bottleneck for developers-not only disrupting the workflow and reducing productivity, but also diverting attention from higher-level semantic design. To this end, this paper presents an automated approach that leverages Large Language Models (LLMs) to repair syntax errors in DCS models using a well-designed, knowledge-informed prompting strategy. Specifically, the prompting is derived from a systematic empirical study of common error patterns, identified through expert interviews and student workshops. It equips the LLM with DCS-specific domain knowledge, including formal grammar rules and illustrative examples, to guide accurate corrections. To evaluate our method, we constructed a new benchmark by systematically injecting realistic syntax errors into validated DCS models. The quantitative evaluation demonstrates the high effectiveness of the proposed approach in terms of repair accuracy and its practical utility regarding time, achieving a speedup of 3.46 times compared to human developers. The experimental replication suite, including the benchmark and prompts, is available at https://github.com/Uuusay1432/DCSModelRepair.git

Paperid: 52, https://arxiv.org/pdf/2512.07237.pdf GitHub GitHub

Abstract:
Transformers have emerged as a universal backbone across 3D perception, video generation, and world models for autonomous driving and embodied AI, where understanding camera geometry is essential for grounding visual observations in three-dimensional space. However, existing camera encoding methods often rely on simplified pinhole assumptions, restricting generalization across the diverse intrinsics and lens distortions in real-world cameras. We introduce Relative Ray Encoding, a geometry-consistent representation that unifies complete camera information, including 6-DoF poses, intrinsics, and lens distortions. To evaluate its capability under diverse controllability demands, we adopt camera-controlled text-to-video generation as a testbed task. Within this setting, we further identify pitch and roll as two components effective for Absolute Orientation Encoding, enabling full control over the initial camera orientation. Together, these designs form UCPE (Unified Camera Positional Encoding), which integrates into a pretrained video Diffusion Transformer through a lightweight spatial attention adapter, adding less than 1% trainable parameters while achieving state-of-the-art camera controllability and visual fidelity. To facilitate systematic training and evaluation, we construct a large video dataset covering a wide range of camera motions and lens types. Extensive experiments validate the effectiveness of UCPE in camera-controllable video generation and highlight its potential as a general camera representation for Transformers across future multi-view, video, and 3D tasks. Code will be available at https://github.com/chengzhag/UCPE.

Paperid: 53, https://arxiv.org/pdf/2512.07232.pdf GitHub

Abstract:
Product matching aims to identify identical or similar products sold on different platforms. By building knowledge graphs (KGs), the product matching problem can be converted to the Entity Alignment (EA) task, which aims to discover the equivalent entities from diverse KGs. The existing EA methods inadequately utilize both attribute triples and relation triples simultaneously, especially the interactions between them. This paper introduces a two-stage pipeline consisting of rough filter and fine filter to match products from eBay and Amazon. For fine filtering, a new framework for Entity Alignment, Relation-aware and Attribute-aware Graph Attention Networks for Entity Alignment (RAEA), is employed. RAEA focuses on the interactions between attribute triples and relation triples, where the entity representation aggregates the alignment signals from attributes and relations with Attribute-aware Entity Encoder and Relation-aware Graph Attention Networks. The experimental results indicate that the RAEA model achieves significant improvements over 12 baselines on EA task in the cross-lingual dataset DBP15K (6.59% on average Hits@1) and delivers competitive results in the monolingual dataset DWY100K. The source code for experiments on DBP15K and DWY100K is available at github (https://github.com/Mockingjay-liu/RAEA-model-for-Entity-Alignment).

Paperid: 54, https://arxiv.org/pdf/2512.07229.pdf GitHub

Abstract:
Generalized Category Discovery (GCD) faces the challenge of categorizing unlabeled data containing both known and novel classes, given only labels for known classes. Previous studies often treat each class independently, neglecting the inherent inter-class relations. Obtaining such inter-class relations directly presents a significant challenge in real-world scenarios. To address this issue, we propose ReLKD, an end-to-end framework that effectively exploits implicit inter-class relations and leverages this knowledge to enhance the classification of novel classes. ReLKD comprises three key modules: a target-grained module for learning discriminative representations, a coarse-grained module for capturing hierarchical class relations, and a distillation module for transferring knowledge from the coarse-grained module to refine the target-grained module's representation learning. Extensive experiments on four datasets demonstrate the effectiveness of ReLKD, particularly in scenarios with limited labeled data. The code for ReLKD is available at https://github.com/ZhouF-ECNU/ReLKD.

Paperid: 55, https://arxiv.org/pdf/2512.07222.pdf GitHub

Abstract:
To address the trade-off between robustness and performance for robust VLM, we observe that function words could incur vulnerability of VLMs against cross-modal adversarial attacks, and propose Function-word De-Attention (FDA) accordingly to mitigate the impact of function words. Similar to differential amplifiers, our FDA calculates the original and the function-word cross-attention within attention heads, and differentially subtracts the latter from the former for more aligned and robust VLMs. Comprehensive experiments include 2 SOTA baselines under 6 different attacks on 2 downstream tasks, 3 datasets, and 3 models. Overall, our FDA yields an average 18/13/53% ASR drop with only 0.2/0.3/0.6% performance drops on the 3 tested models on retrieval, and a 90% ASR drop with a 0.3% performance gain on visual grounding. We demonstrate the scalability, generalization, and zero-shot performance of FDA experimentally, as well as in-depth ablation studies and analysis. Code will be made publicly at https://github.com/michaeltian108/FDA.

Paperid: 56, https://arxiv.org/pdf/2512.07221.pdf GitHub

Abstract:
Simultaneous localization and mapping (SLAM) plays a fundamental role in extended reality (XR) applications. As the standards for immersion in XR continue to increase, the demands for SLAM benchmarking have become more stringent. Trajectory accuracy is the key metric, and marker-based optical motion capture (MoCap) systems are widely used to generate ground truth (GT) because of their drift-free and relatively accurate measurements. However, the precision of MoCap-based GT is limited by two factors: the spatiotemporal calibration with the device under test (DUT) and the inherent jitter in the MoCap measurements. These limitations hinder accurate SLAM benchmarking, particularly for key metrics like rotation error and inter-frame jitter, which are critical for immersive XR experiences. This paper presents a novel continuous-time maximum likelihood estimator to address these challenges. The proposed method integrates auxiliary inertial measurement unit (IMU) data to compensate for MoCap jitter. Additionally, a variable time synchronization method and a pose residual based on screw congruence constraints are proposed, enabling precise spatiotemporal calibration across multiple sensors and the DUT. Experimental results demonstrate that our approach outperforms existing methods, achieving the precision necessary for comprehensive benchmarking of state-of-the-art SLAM algorithms in XR applications. Furthermore, we thoroughly validate the practicality of our method by benchmarking several leading XR devices and open-source SLAM algorithms. The code is publicly available at https://github.com/ylab-xrpg/xr-hpgt.

Paperid: 57, https://arxiv.org/pdf/2512.07201.pdf GitHub

Abstract:
Diffusion models have achieved remarkable performance in generative modeling, yet their theoretical foundations are often intricate, and the gap between mathematical formulations in papers and practical open-source implementations can be difficult to bridge. Existing tutorials primarily focus on deriving equations, offering limited guidance on how diffusion models actually operate in code. To address this, we present a concise implementation of approximately 300 lines that explains diffusion models from a code-execution perspective. Our minimal example preserves the essential components -- including forward diffusion, reverse sampling, the noise-prediction network, and the training loop -- while removing unnecessary engineering details. This technical report aims to provide researchers with a clear, implementation-first understanding of how diffusion models work in practice and how code and theory correspond. Our code and pre-trained models are available at: https://github.com/disanda/GM/tree/main/DDPM-DDIM-ClassifierFree.

Paperid: 58, https://arxiv.org/pdf/2512.07200.pdf GitHub

Abstract:
In bus arrival time prediction, the process of organizing road infrastructure network data into homogeneous entities is known as segmentation. Segmenting a road network is widely recognized as the first and most critical step in developing an arrival time prediction system, particularly for auto-regressive-based approaches. Traditional methods typically employ a uniform segmentation strategy, which fails to account for varying physical constraints along roads, such as road conditions, intersections, and points of interest, thereby limiting prediction efficiency. In this paper, we propose a Reinforcement Learning (RL)-based approach to efficiently and adaptively learn non-uniform road segments for arrival time prediction. Our method decouples the prediction process into two stages: 1) Non-uniform road segments are extracted based on their impact scores using the proposed RL framework; and 2) A linear prediction model is applied to the selected segments to make predictions. This method ensures optimal segment selection while maintaining computational efficiency, offering a significant improvement over traditional uniform approaches. Furthermore, our experimental results suggest that the linear approach can even achieve better performance than more complex methods. Extensive experiments demonstrate the superiority of the proposed method, which not only enhances efficiency but also improves learning performance on large-scale benchmarks. The dataset and the code are publicly accessible at: https://github.com/pangjunbiao/Less-is-More.

Paperid: 59, https://arxiv.org/pdf/2512.07198.pdf GitHub

Abstract:
An image can convey a compelling story by presenting rich, logically connected visual clues. These connections form Chains-of-Reasoning (CoRs) within the image, enabling viewers to infer events, causal relationships, and other information, thereby understanding the underlying story. In this paper, we focus on these semantically rich images and define them as Storytelling Images. Such images have diverse applications beyond illustration creation and cognitive screening, leveraging their ability to convey multi-layered information visually and inspire active interpretation. However, due to their complex semantic nature, Storytelling Images are inherently challenging to create, and thus remain relatively scarce. To address this challenge, we introduce the Storytelling Image Generation task, which explores how generative AI models can be leveraged to create such images. Specifically, we propose a two-stage pipeline, StorytellingPainter, which combines the creative reasoning abilities of Large Language Models (LLMs) with the visual synthesis capabilities of Text-to-Image (T2I) models to generate Storytelling Images. Alongside this pipeline, we develop a dedicated evaluation framework comprising three main evaluators: a Semantic Complexity Evaluator, a KNN-based Diversity Evaluator and a Story-Image Alignment Evaluator. Given the critical role of story generation in the Storytelling Image Generation task and the performance disparity between open-source and proprietary LLMs, we further explore tailored training strategies to reduce this gap, resulting in a series of lightweight yet effective models named Mini-Storytellers. Experimental results demonstrate the feasibility and effectiveness of our approaches. The code is available at https://github.com/xiujiesong/StorytellingImageGeneration.

Paperid: 60, https://arxiv.org/pdf/2512.07197.pdf GitHub

Abstract:
3D Gaussian Splatting (3DGS) has emerged as a powerful explicit representation enabling real-time, high-fidelity 3D reconstruction and novel view synthesis. However, its practical use is hindered by the massive memory and computational demands required to store and render millions of Gaussians. These challenges become even more severe in 4D dynamic scenes. To address these issues, the field of Efficient Gaussian Splatting has rapidly evolved, proposing methods that reduce redundancy while preserving reconstruction quality. This survey provides the first unified overview of efficient 3D and 4D Gaussian Splatting techniques. For both 3D and 4D settings, we systematically categorize existing methods into two major directions, Parameter Compression and Restructuring Compression, and comprehensively summarize the core ideas and methodological trends within each category. We further cover widely used datasets, evaluation metrics, and representative benchmark comparisons. Finally, we discuss current limitations and outline promising research directions toward scalable, compact, and real-time Gaussian Splatting for both static and dynamic 3D scene representation.

Paperid: 61, https://arxiv.org/pdf/2512.07194.pdf GitHub

Abstract:
While surrogate backpropagation proves useful for training deep spiking neural networks (SNNs), incorporating biologically inspired local signals on a large scale remains challenging. This difficulty stems primarily from the high memory demands of maintaining accurate spike-timing logs and the potential for purely local plasticity adjustments to clash with the supervised learning goal. To effectively leverage local signals derived from spiking neuron dynamics, we introduce Dopamine-Modulated Spike-Synchrony-Dependent Plasticity (DA-SSDP), a synchrony-based rule that is sensitive to loss and brings a synchrony-based local learning signal to the model. DA-SSDP condenses spike patterns into a synchrony metric at the batch level. An initial brief warm-up phase assesses its relationship to the task loss and sets a fixed gate that subsequently adjusts the local update's magnitude. In cases where synchrony proves unrelated to the task, the gate settles at one, simplifying DA-SSDP to a basic two-factor synchrony mechanism that delivers minor weight adjustments driven by concurrent spike firing and a Gaussian latency function. These small weight updates are only added to the network`s deeper layers following the backpropagation phase, and our tests showed this simplified version did not degrade performance and sometimes gave a small accuracy boost, serving as a regularizer during training. The rule stores only binary spike indicators and first-spike latencies with a Gaussian kernel. Without altering the model structure or optimization routine, evaluations on benchmarks like CIFAR-10 (+0.42\%), CIFAR-100 (+0.99\%), CIFAR10-DVS (+0.1\%), and ImageNet-1K (+0.73\%) demonstrated consistent accuracy gains, accompanied by a minor increase in computational overhead. Our code is available at https://github.com/NeuroSyd/DA-SSDP.

Paperid: 62, https://arxiv.org/pdf/2512.07193.pdf GitHub

Abstract:
This paper investigates the semantic robustness of attention-based classifiers for design pattern detection, particularly focusing on their reliance on structural and behavioral semantics. We reproduce the DPDAtt, an attention-based design pattern detection approach using learning-based classifiers, and evaluate its performance under obfuscation. To this end, we curate an obfuscated version of the DPDAtt Corpus, where the name identifiers in code such as class names, method names, etc., and string literals like print statements and comment blocks are replaced while preserving control flow, inheritance, and logic. Our findings reveal that these trained classifiers in DPDAtt depend significantly on superficial syntactic features, leading to substantial misclassification when such cues are removed through obfuscation. This work highlights the need for more robust detection tools capable of capturing deeper semantic meanings in source code. We propose our curated Obfuscated corpus (containing 34 Java source files) as a reusable proof-of-concept benchmark for evaluating state-of-the-art design pattern detectors on their true semantic generalization capabilities.

Paperid: 63, https://arxiv.org/pdf/2512.07171.pdf GitHub

Abstract:
Underwater image restoration is essential for marine applications ranging from ecological monitoring to archaeological surveys, but effectively addressing the complex and spatially varying nature of underwater degradations remains a challenge. Existing methods typically apply uniform restoration strategies across the entire image, struggling to handle multiple co-occurring degradations that vary spatially and with water conditions. We introduce TIDE, a $\underline{t}$wo stage $\underline{i}$nverse $\underline{d}$egradation $\underline{e}$stimation framework that explicitly models degradation characteristics and applies targeted restoration through specialized prior decomposition. Our approach disentangles the restoration process into multiple specialized hypotheses that are adaptively fused based on local degradation patterns, followed by a progressive refinement stage that corrects residual artifacts. Specifically, TIDE decomposes underwater degradations into four key factors, namely color distortion, haze, detail loss, and noise, and designs restoration experts specialized for each. By generating specialized restoration hypotheses, TIDE balances competing degradation factors and produces natural results even in highly degraded regions. Extensive experiments across both standard benchmarks and challenging turbid water conditions show that TIDE achieves competitive performance on reference based fidelity metrics while outperforming state of the art methods on non reference perceptual quality metrics, with strong improvements in color correction and contrast enhancement. Our code is available at: https://rakesh-123-cryp.github.io/TIDE.

Paperid: 64, https://arxiv.org/pdf/2512.07155.pdf GitHub

Abstract:
Diffusion models exhibit remarkable generative ability, yet achieving smooth and semantically consistent image morphing remains a challenge. Existing approaches often yield abrupt transitions or over-saturated appearances due to the lack of adaptive structural and semantic alignments. We propose CHIMERA, a zero-shot diffusion-based framework that formulates morphing as a cached inversion-guided denoising process. To handle large semantic and appearance disparities, we propose Adaptive Cache Injection and Semantic Anchor Prompting. Adaptive Cache Injection (ACI) caches down, mid, and up blocks features from both inputs during DDIM inversion and re-injects them adaptively during denoising, enabling spatial and semantic alignment in depth- and time-adaptive manners and enabling natural feature fusion and smooth transitions. Semantic Anchor Prompting (SAP) leverages a vision-language model to generate a shared anchor prompt that serves as a semantic anchor, bridging dissimilar inputs and guiding the denoising process toward coherent results. Finally, we introduce the Global-Local Consistency Score (GLCS), a morphing-oriented metric that simultaneously evaluates the global harmonization of the two inputs and the smoothness of the local morphing transition. Extensive experiments and user studies show that CHIMERA achieves smoother and more semantically aligned transitions than existing methods, establishing a new state of the art in image morphing. The code and project page will be publicly released.

Paperid: 65, https://arxiv.org/pdf/2512.07141.pdf GitHub

Abstract:
As multimodal reasoning improves the overall capabilities of Large Vision Language Models (LVLMs), recent studies have begun to explore safety-oriented reasoning, aiming to enhance safety awareness by analyzing potential safety risks during the reasoning process before generating the final response. Although such approaches improve safety awareness and interpretability, this single-pass think-then-answer paradigm remains vulnerable to contextual or visual jailbreak attacks. This reveals a critical flaw: single-pass reasoning may overlook explicit harmful content in its own output. Our key insight is to exploit this wasted signal through reflection, which can effectively leverage the malicious content revealed in the first-pass reasoning to enable genuine self-correction and prevent unsafe generations. Motivated by this, we propose Think-Reflect-Revise (TRR), a three-stage training framework designed to enhance the safety alignment of LVLMs through policy-guided self-reflection. We first build a Reflective Safety Reasoning (ReSafe) dataset with 5,000 examples that follow a think-reflect-revise process. We then fine-tune the target model using the ReSafe dataset to initialize reflective behavior, and finally reinforce policy-guided reflection through reinforcement learning. Experimental results show that TRR substantially improves the safety performance of LVLMs across both safety-awareness benchmarks and jailbreak attack evaluations, increasing the overall safe response rate from 42.8% to 87.7% on Qwen2.5-VL-7B, while preserving stable performance on general benchmarks such as MMMU and MMStar. The project page is available at https://think-reflect-revise.github.io/.

Paperid: 66, https://arxiv.org/pdf/2512.07136.pdf GitHub GitHub

Abstract:
Multimodal human action recognition (HAR) leverages complementary sensors for activity classification. Beyond recognition, recent advances in large language models (LLMs) enable detailed descriptions and causal reasoning, motivating new tasks: human action understanding (HAU) and human action reasoning (HARn). However, most LLMs, especially large vision language models (LVLMs), struggle with non-RGB modalities such as depth, IMU, and mmWave due to the lack of large-scale data-caption resources. Existing HAR datasets mainly provide coarse data-label annotations, which are insufficient to capture fine-grained action dynamics needed for HAU and HARn. We consider two ground-truth pair types: (1) data label (discrete category) and (2) data caption (textual description). Naively generating captions from labels often lacks logical and spatiotemporal consistency. We introduce CUHK-X, a large-scale multimodal dataset and benchmark suite for HAR, HAU, and HARn. CUHK-X contains 58,445 samples covering 40 actions performed by 30 participants across two indoor environments. To improve caption consistency, we propose a prompt-based scene creation method that leverages LLMs to generate logically connected activity sequences, followed by human validation. CUHK-X includes three benchmarks with six evaluation tasks. Experiments report average accuracies of 76.52% (HAR), 40.76% (HAU), and 70.25% (HARn). CUHK-X aims to enable the community to apply and develop data-intensive learning methods for robust, multimodal human activity analysis. Project page and code: https://openaiotlab.github.io/CUHK-X/ and https://github.com/openaiotlab/CUHK-X.

Paperid: 67, https://arxiv.org/pdf/2512.07132.pdf GitHub

Abstract:
Specialized visual tools can augment large language models or vision language models with expert knowledge (e.g., grounding, spatial reasoning, medical knowledge, etc.), but knowing which tools to call (and when to call them) can be challenging. We introduce DART, a multi-agent framework that uses disagreements between multiple debating visual agents to identify useful visual tools (e.g., object detection, OCR, spatial reasoning, etc.) that can resolve inter-agent disagreement. These tools allow for fruitful multi-agent discussion by introducing new information, and by providing tool-aligned agreement scores that highlight agents in agreement with expert tools, thereby facilitating discussion. We utilize an aggregator agent to select the best answer by providing the agent outputs and tool information. We test DART on four diverse benchmarks and show that our approach improves over multi-agent debate as well as over single agent tool-calling frameworks, beating the next-strongest baseline (multi-agent debate with a judge model) by 3.4% and 2.4% on A-OKVQA and MMMU respectively. We also find that DART adapts well to new tools in applied domains, with a 1.3% improvement on the M3D medical dataset over other strong tool-calling, single agent, and multi-agent baselines. Additionally, we measure text overlap across rounds to highlight the rich discussion in DART compared to existing multi-agent methods. Finally, we study the tool call distribution, finding that diverse tools are reliably used to help resolve disagreement.

Paperid: 68, https://arxiv.org/pdf/2512.07130.pdf GitHub

Abstract:
End-to-end autonomous driving has emerged as a pivotal direction in the field of autonomous systems. Recent works have demonstrated impressive performance by incorporating high-level guidance signals to steer low-level trajectory planners. However, their potential is often constrained by inaccurate high-level guidance and the computational overhead of complex guidance modules. To address these limitations, we propose Mimir, a novel hierarchical dual-system framework capable of generating robust trajectories relying on goal points with uncertainty estimation: (1) Unlike previous approaches that deterministically model, we estimate goal point uncertainty with a Laplace distribution to enhance robustness; (2) To overcome the slow inference speed of the guidance system, we introduce a multi-rate guidance mechanism that predicts extended goal points in advance. Validated on challenging Navhard and Navtest benchmarks, Mimir surpasses previous state-of-the-art methods with a 20% improvement in the driving score EPDMS, while achieving 1.6 times improvement in high-level module inference speed without compromising accuracy. The code and models will be released soon to promote reproducibility and further development. The code is available at https://github.com/ZebinX/Mimir-Uncertainty-Driving

Paperid: 69, https://arxiv.org/pdf/2512.07126.pdf GitHub

Abstract:
VTON (Virtual Try-ON) aims at synthesizing the target clothing on a certain person, preserving the details of the target clothing while keeping the rest of the person unchanged. Existing methods suffer from the discrepancies between the generated clothing results and the target ones, in terms of the patterns, textures and boundaries. Therefore, we propose to use an energy function to impose constraints on the attention map extracted through the generation process. Thus, at each generation step, the attention can be more focused on the clothing region of interest, thereby influencing the generation results to be more consistent with the target clothing details. Furthermore, to address the limitation that existing evaluation metrics concentrate solely on image realism and overlook the alignment with target elements, we design a new metric, Virtual Try-on Inception Distance (VTID), to bridge this gap and ensure a more comprehensive assessment. On the VITON-HD and DressCode datasets, our approach has outperformed the previous state-of-the-art (SOTA) methods by 1.4%, 2.3%, 12.3%, and 5.8% in the traditional metrics of LPIPS, FID, KID, and the new VTID metrics, respectively. Additionally, by applying the generated data to downstream Clothing-Change Re-identification (CC-Reid) methods, we have achieved performance improvements of 2.5%, 1.1%, and 1.6% on the LTCC, PRCC, VC-Clothes datasets in the metrics of Rank-1. The code of our method is public at https://github.com/MrWhiteSmall/CSC-VTON.git.

Paperid: 70, https://arxiv.org/pdf/2512.07113.pdf GitHub

Abstract:
Understanding the underlying linguistic rules of plant genomes remains a fundamental challenge in computational biology. Recent advances including AgroNT and PDLLMs have made notable progress although, they suffer from excessive parameter size and limited ability to model the bidirectional nature of DNA strands respectively. To address these limitations, we propose PlantBiMoE, a lightweight and expressive plant genome language model that integrates bidirectional Mamba and a Sparse Mixture-of-Experts (SparseMoE) framework. The bidirectional Mamba enables the model to effectively capture structural dependencies across both the forward and reverse DNA strands, while SparseMoE significantly reduces the number of active parameters, improving computational efficiency without sacrificing modeling capacity. We evaluated and tested our model on the Modified Plants Genome Benchmark (MPGB), an enhanced genomic benchmark, which consolidates 31 datasets across 11 representative tasks, with input sequence lengths ranging from 50 to 6,000 bp. Experimental results demonstrate that PlantBiMoE achieves the best performance on 20 out of 31 datasets and the average best when comparing with existing models. In summary, all above results demonstrate that our model can effectively represent plant genomic sequences, serving as a robust computational tool for diverse genomic tasks, while making substantive contributions to plant genomics, gene editing, and synthetic biology. The code is available at: https://github.com/HUST-Keep-Lin/PlantBiMoE

Paperid: 71, https://arxiv.org/pdf/2512.07107.pdf GitHub

Abstract:
We present COREA, the first unified framework that jointly learns relightable 3D Gaussians and a Signed Distance Field (SDF) for accurate geometry reconstruction and faithful relighting. While recent 3D Gaussian Splatting (3DGS) methods have extended toward mesh reconstruction and physically-based rendering (PBR), their geometry is still learned from 2D renderings, leading to coarse surfaces and unreliable BRDF-lighting decomposition. To address these limitations, COREA introduces a coarse-to-fine bidirectional 3D-to-3D alignment strategy that allows geometric signals to be learned directly in 3D space. Within this strategy, depth provides coarse alignment between the two representations, while depth gradients and normals refine fine-scale structure, and the resulting geometry supports stable BRDF-lighting decomposition. A density-control mechanism further stabilizes Gaussian growth, balancing geometric fidelity with memory efficiency. Experiments on standard benchmarks demonstrate that COREA achieves superior performance in novel-view synthesis, mesh reconstruction, and PBR within a unified framework.

Paperid: 72, https://arxiv.org/pdf/2512.07079.pdf GitHub

Abstract:
Progress in computer-aided synthesis planning (CASP) is obscured by the lack of standardized evaluation infrastructure and the reliance on metrics that prioritize topological completion over chemical validity. We introduce RetroCast, a unified evaluation suite that standardizes heterogeneous model outputs into a common schema to enable statistically rigorous, apples-to-apples comparison. The framework includes a reproducible benchmarking pipeline with stratified sampling and bootstrapped confidence intervals, accompanied by SynthArena, an interactive platform for qualitative route inspection. We utilize this infrastructure to evaluate leading search-based and sequence-based algorithms on a new suite of standardized benchmarks. Our analysis reveals a divergence between "solvability" (stock-termination rate) and route quality; high solvability scores often mask chemical invalidity or fail to correlate with the reproduction of experimental ground truths. Furthermore, we identify a "complexity cliff" in which search-based methods, despite high solvability rates, exhibit a sharp performance decay in reconstructing long-range synthetic plans compared to sequence-based approaches. We release the full framework, benchmark definitions, and a standardized database of model predictions to support transparent and reproducible development in the field.

Paperid: 73, https://arxiv.org/pdf/2512.07076.pdf GitHub

Abstract:
Camouflage is primarily context-dependent yet current metrics for camouflaged scenarios overlook this critical factor. Instead, these metrics are originally designed for evaluating general or salient objects, with an inherent assumption of uncorrelated spatial context. In this paper, we propose a new contextualized evaluation paradigm, Context-measure, built upon a probabilistic pixel-aware correlation framework. By incorporating spatial dependencies and pixel-wise camouflage quantification, our measure better aligns with human perception. Extensive experiments across three challenging camouflaged object segmentation datasets show that Context-measure delivers more reliability than existing context-independent metrics. Our measure can provide a foundational evaluation benchmark for various computer vision applications involving camouflaged patterns, such as agricultural, industrial, and medical scenarios. Code is available at https://github.com/pursuitxi/Context-measure.

Paperid: 74, https://arxiv.org/pdf/2512.07062.pdf GitHub

Abstract:
Although diffusion models with strong visual priors have emerged as powerful dense prediction backboens, they overlook a core limitation: the stochastic noise at the core of diffusion sampling is inherently misaligned with dense prediction that requires a deterministic mapping from image to geometry. In this paper, we show that this stochastic noise corrupts fine-grained spatial cues and pushes the model toward timestep-specific noise objectives, consequently destroying meaningful geometric structure mappings. To address this, we introduce $\mathrm{D}^{\mathrm{3}}$-Predictor, a noise-free deterministic framework built by reformulating a pretrained diffusion model without stochasticity noise. Instead of relying on noisy inputs to leverage diffusion priors, $\mathrm{D}^{\mathrm{3}}$-Predictor views the pretrained diffusion network as an ensemble of timestep-dependent visual experts and self-supervisedly aggregates their heterogeneous priors into a single, clean, and complete geometric prior. Meanwhile, we utilize task-specific supervision to seamlessly adapt this noise-free prior to dense prediction tasks. Extensive experiments on various dense prediction tasks demonstrate that $\mathrm{D}^{\mathrm{3}}$-Predictor achieves competitive or state-of-the-art performance in diverse scenarios. In addition, it requires less than half the training data previously used and efficiently performs inference in a single step. Our code, data, and checkpoints are publicly available at https://x-gengroup.github.io/HomePage_D3-Predictor/.

Paperid: 75, https://arxiv.org/pdf/2512.07059.pdf GitHub

Abstract:
Despite substantial investment in safety alignment, the vulnerability of large language models to sophisticated multi-turn adversarial attacks remains poorly characterized, and whether model scale or inference mode affects robustness is unknown. This study employed the TEMPEST multi-turn attack framework to evaluate ten frontier models from eight vendors across 1,000 harmful behaviors, generating over 97,000 API queries across adversarial conversations with automated evaluation by independent safety classifiers. Results demonstrated a spectrum of vulnerability: six models achieved 96% to 100% attack success rate (ASR), while four showed meaningful resistance, with ASR ranging from 42% to 78%; enabling extended reasoning on identical architecture reduced ASR from 97% to 42%. These findings indicate that safety alignment quality varies substantially across vendors, that model scale does not predict adversarial robustness, and that thinking mode provides a deployable safety enhancement. Collectively, this work establishes that current alignment techniques remain fundamentally vulnerable to adaptive multi-turn attacks regardless of model scale, while identifying deliberative inference as a promising defense direction.

Paperid: 76, https://arxiv.org/pdf/2512.07019.pdf GitHub

Abstract:
The proliferation of Large Language Models (LLMs) necessitates valid evaluation methods to provide guidance for both downstream applications and actionable future improvements. The Item Response Theory (IRT) model with Computerized Adaptive Testing has recently emerged as a promising framework for evaluating LLMs via their response accuracy. Beyond simple response accuracy, LLMs' chain of thought (CoT) lengths serve as a vital indicator of their reasoning ability. To leverage the CoT length information to assist the evaluation of LLMs, we propose the Latency-Response Theory (LaRT) model, which jointly models both the response accuracy and CoT length by introducing a key correlation parameter between the latent ability and the latent speed. We derive an efficient stochastic approximation Expectation-Maximization algorithm for parameter estimation. We establish rigorous identifiability results for the latent ability and latent speed parameters to ensure the statistical validity of their estimation. Through both theoretical asymptotic analyses and simulation studies, we demonstrate LaRT's advantages over IRT in terms of superior estimation accuracy and shorter confidence intervals for latent trait estimation. To evaluate LaRT in real data, we collect responses from diverse LLMs on popular benchmark datasets. We find that LaRT yields different LLM rankings than IRT and outperforms IRT across multiple key evaluation metrics including predictive power, item efficiency, ranking validity, and LLM evaluation efficiency. Code and data are available at https://github.com/Toby-X/Latency-Response-Theory-Model.

Paperid: 77, https://arxiv.org/pdf/2512.07011.pdf GitHub

Abstract:
Modern large language models increasingly require long contexts for reasoning and multi-document tasks, but attention's quadratic complexity creates a severe computational bottleneck. We present Block-Sparse FlashAttention (BSFA), a drop-in replacement that accelerates long-context inference while preserving model quality. Unlike methods that predict importance before computing scores, BSFA computes exact query-key similarities to select the top-k most important value blocks for each query. By comparing per-block maximum scores against calibrated thresholds, we skip approximately 50% of the computation and memory transfers for pruned blocks. Our training-free approach requires only a one-time threshold calibration on a small dataset to learn the per-layer and per-head attention score distributions. We provide a CUDA kernel implementation that can be used as a drop-in replacement for FlashAttention. On Llama-3.1-8B, BSFA achieves up to 1.10x speedup on real-world reasoning benchmarks and up to 1.24x for needle-in-a-haystack retrieval tasks while maintaining above 99% baseline accuracy, with certain configurations even improving accuracy by focusing on the most relevant content, substantially outperforming existing sparse attention methods. The implementation is available at https://github.com/Danielohayon/Block-Sparse-Flash-Attention

Paperid: 78, https://arxiv.org/pdf/2512.06976.pdf GitHub

Abstract:
This paper presents a novel post-processing methodology for extracting high-quality geometries from density-based topology optimization results. Current post-processing approaches often struggle to simultaneously achieve smooth boundaries, preserve volume fraction, and maintain topological features. We propose a robust method based on a signed distance function (SDF) that addresses these challenges through a two-stage process: first, an SDF representation of density isocontours is constructed, which is followed by geometry refinement using radial basis functions (RBFs). The method generates smooth boundary representations that appear to originate from much finer discretizations while maintaining the computational efficiency of coarse mesh optimization. Through comprehensive validation, our approach demonstrates a 18% reduction in maximum equivalent stress values compared to conventional methods, achieved through continuous geometric transitions at boundaries. The resulting implicit boundary representation facilitates seamless export to standard manufacturing formats without intermediate reconstruction steps, providing a robust foundation for practical engineering applications where high-quality geometric representations are essential.

Paperid: 79, https://arxiv.org/pdf/2512.06949.pdf GitHub

Abstract:
Histopathology image segmentation is essential for delineating tissue structures in skin cancer diagnostics, but modeling spatial context and inter-tissue relationships remains a challenge, especially in regions with overlapping or morphologically similar tissues. Current convolutional neural network (CNN)-based approaches operate primarily on visual texture, often treating tissues as independent regions and failing to encode biological context. To this end, we introduce Neural Tissue Relation Modeling (NTRM), a novel segmentation framework that augments CNNs with a tissue-level graph neural network to model spatial and functional relationships across tissue types. NTRM constructs a graph over predicted regions, propagates contextual information via message passing, and refines segmentation through spatial projection. Unlike prior methods, NTRM explicitly encodes inter-tissue dependencies, enabling structurally coherent predictions in boundary-dense zones. On the benchmark Histopathology Non-Melanoma Skin Cancer Segmentation Dataset, NTRM outperforms state-of-the-art methods, achieving a robust Dice similarity coefficient that is 4.9\% to 31.25\% higher than the best-performing models among the evaluated approaches. Our experiments indicate that relational modeling offers a principled path toward more context-aware and interpretable histological segmentation, compared to local receptive-field architectures that lack tissue-level structural awareness. Our code is available at https://github.com/shravan-18/NTRM.

Paperid: 80, https://arxiv.org/pdf/2512.06905.pdf GitHub

Abstract:
Reference-to-video (R2V) generation aims to synthesize videos that align with a text prompt while preserving the subject identity from reference images. However, current R2V methods are hindered by the reliance on explicit reference image-video-text triplets, whose construction is highly expensive and difficult to scale. We bypass this bottleneck by introducing Saber, a scalable zero-shot framework that requires no explicit R2V data. Trained exclusively on video-text pairs, Saber employs a masked training strategy and a tailored attention-based model design to learn identity-consistent and reference-aware representations. Mask augmentation techniques are further integrated to mitigate copy-paste artifacts common in reference-to-video generation. Moreover, Saber demonstrates remarkable generalization capabilities across a varying number of references and achieves superior performance on the OpenS2V-Eval benchmark compared to methods trained with R2V data.

Paperid: 81, https://arxiv.org/pdf/2512.06899.pdf GitHub

Abstract:
Transferable backdoors pose a severe threat to the Pre-trained Language Models (PLMs) supply chain, yet defensive research remains nascent, primarily relying on detecting anomalies in the output feature space. We identify a critical flaw that fine-tuning on downstream tasks inevitably modifies model parameters, shifting the output distribution and rendering pre-computed defense ineffective. To address this, we propose Patronus, a novel framework that use input-side invariance of triggers against parameter shifts. To overcome the convergence challenges of discrete text optimization, Patronus introduces a multi-trigger contrastive search algorithm that effectively bridges gradient-based optimization with contrastive learning objectives. Furthermore, we employ a dual-stage mitigation strategy combining real-time input monitoring with model purification via adversarial training. Extensive experiments across 15 PLMs and 10 tasks demonstrate that Patronus achieves $\geq98.7\%$ backdoor detection recall and reduce attack success rates to clean settings, significantly outperforming all state-of-the-art baselines in all settings. Code is available at https://github.com/zth855/Patronus.

Paperid: 82, https://arxiv.org/pdf/2512.06886.pdf GitHub

Abstract:
Unsupervised domain adaptation (UDA) for semantic segmentation aims to transfer knowledge from a labeled source domain to an unlabeled target domain. Despite the effectiveness of self-training techniques in UDA, they struggle to learn each class in a balanced manner due to inherent class imbalance and distribution shift in both data and label space between domains. To address this issue, we propose Balanced Learning for Domain Adaptation (BLDA), a novel approach to directly assess and alleviate class bias without requiring prior knowledge about the distribution shift. First, we identify over-predicted and under-predicted classes by analyzing the distribution of predicted logits. Subsequently, we introduce a post-hoc approach to align the logits distributions across different classes using shared anchor distributions. To further consider the network's need to generate unbiased pseudo-labels during self-training, we estimate logits distributions online and incorporate logits correction terms into the loss function. Moreover, we leverage the resulting cumulative density as domain-shared structural knowledge to connect the source and target domains. Extensive experiments on two standard UDA semantic segmentation benchmarks demonstrate that BLDA consistently improves performance, especially for under-predicted classes, when integrated into various existing methods. Code is available at https://github.com/Woof6/BLDA.

Paperid: 83, https://arxiv.org/pdf/2512.06885.pdf GitHub

Abstract:
Panorama generation has recently attracted growing interest in the research community, with two core tasks, text-to-panorama and view-to-panorama generation. However, existing methods still face two major challenges: their U-Net-based architectures constrain the visual quality of the generated panoramas, and they usually treat the two core tasks independently, which leads to modeling redundancy and inefficiency. To overcome these challenges, we propose a joint-face panorama (JoPano) generation approach that unifies the two core tasks within a DiT-based model. To transfer the rich generative capabilities of existing DiT backbones learned from natural images to the panorama domain, we propose a Joint-Face Adapter built on the cubemap representation of panoramas, which enables a pretrained DiT to jointly model and generate different views of a panorama. We further apply Poisson Blending to reduce seam inconsistencies that often appear at the boundaries between cube faces. Correspondingly, we introduce Seam-SSIM and Seam-Sobel metrics to quantitatively evaluate the seam consistency. Moreover, we propose a condition switching mechanism that unifies text-to-panorama and view-to-panorama tasks within a single model. Comprehensive experiments show that JoPano can generate high-quality panoramas for both text-to-panorama and view-to-panorama generation tasks, achieving state-of-the-art performance on FID, CLIP-FID, IS, and CLIP-Score metrics.

Paperid: 84, https://arxiv.org/pdf/2512.06883.pdf GitHub

Abstract:
Multimodal recommendation enhances accuracy by leveraging visual and textual signals, and its success largely depends on learning high-quality cross-modal representations. Recent advances in Large Vision-Language Models (LVLMs) offer unified multimodal representation learning, making them a promising backbone. However, applying LVLMs to recommendation remains challenging due to (i) representation misalignment, where domain gaps between item data and general pre-training lead to unaligned embedding spaces, and (ii) gradient conflicts during fine-tuning, where shared adapters cause interference and a lack of discriminative power. To address this, we propose SDA, a lightweight framework for Structural and Disentangled Adaptation, which integrates two components: Cross-Modal Structural Alignment (CMSA) and Modality-Disentangled Adaptation. CMSA aligns embeddings using intra-modal structures as a soft teacher, while MoDA mitigates gradient conflicts via expertized, gated low-rank paths to disentangle gradient flows. Experiments on three public Amazon datasets show SDA integrates seamlessly with existing multimodal and sequential recommenders, yielding average gains of 6.15% in Hit@10 and 8.64% in NDCG@10. It also achieves up to 12.83% and 18.70% gains on long-tail items with minimal inference overhead. Our code and full experimental results are available at https://github.com/RaoZhongtao/SDA.

Paperid: 85, https://arxiv.org/pdf/2512.06877.pdf GitHub

Abstract:
Remote sensing scene classification plays a key role in Earth observation by enabling the automatic identification of land use and land cover (LULC) patterns from aerial and satellite imagery. Despite recent progress with convolutional neural networks (CNNs) and vision transformers (ViTs), the task remains challenging due to variations in spatial resolution, viewpoint, orientation, and background conditions, which often reduce the generalization ability of existing models. To address these challenges, this paper proposes a lightweight architecture based on the convolutional mixer paradigm. The model alternates between spatial mixing through depthwise convolutions at multiple scales and channel mixing through pointwise operations, enabling efficient extraction of both local and contextual information while keeping the number of parameters and computations low. Extensive experiments were conducted on the AID and EuroSAT benchmarks. The proposed model achieved overall accuracy, average accuracy, and Kappa values of 74.7%, 74.57%, and 73.79 on the AID dataset, and 93.90%, 93.93%, and 93.22 on EuroSAT, respectively. These results demonstrate that the proposed approach provides a good balance between accuracy and efficiency compared with widely used CNN- and transformer-based models. Code will be publicly available on: https://github.com/mqalkhatib/SceneMixer

Paperid: 86, https://arxiv.org/pdf/2512.06870.pdf GitHub

Abstract:
Pseudo-label learning is widely used in semantic segmentation, particularly in label-scarce scenarios such as unsupervised domain adaptation (UDA) and semisupervised learning (SSL). Despite its success, this paradigm can generate erroneous pseudo-labels, which are further amplified during training due to utilization of one-hot encoding. To address this issue, we propose ECOCSeg, a novel perspective for segmentation models that utilizes error-correcting output codes (ECOC) to create a fine-grained encoding for each class. ECOCSeg offers several advantages. First, an ECOC-based classifier is introduced, enabling model to disentangle classes into attributes and handle partial inaccurate bits, improving stability and generalization in pseudo-label learning. Second, a bit-level label denoising mechanism is developed to generate higher-quality pseudo-labels, providing adequate and robust supervision for unlabeled images. ECOCSeg can be easily integrated with existing methods and consistently demonstrates significant improvements on multiple UDA and SSL benchmarks across different segmentation architectures. Code is available at https://github.com/Woof6/ECOCSeg.

Paperid: 87, https://arxiv.org/pdf/2512.06866.pdf GitHub

Abstract:
Recent advances in Video Large Language Models (VLLMs) have achieved remarkable video understanding capabilities, yet face critical efficiency bottlenecks due to quadratic computational growth with lengthy visual token sequences of long videos. While existing keyframe sampling methods can improve temporal modeling efficiency, additional computational cost is introduced before feature encoding, and the binary frame selection paradigm is found suboptimal. Therefore, in this work, we propose Dynamic Token compression via LLM-guided Keyframe prior (DyToK), a training-free paradigm that enables dynamic token compression by harnessing VLLMs' inherent attention mechanisms. Our analysis reveals that VLLM attention layers naturally encoding query-conditioned keyframe priors, by which DyToK dynamically adjusts per-frame token retention ratios, prioritizing semantically rich frames while suppressing redundancies. Extensive experiments demonstrate that DyToK achieves state-of-the-art efficiency-accuracy tradeoffs. DyToK shows plug-and-play compatibility with existing compression methods, such as VisionZip and FastV, attaining 4.3x faster inference while preserving accuracy across multiple VLLMs, such as LLaVA-OneVision and Qwen2.5-VL. Code is available at https://github.com/yu-lin-li/DyToK .

Paperid: 88, https://arxiv.org/pdf/2512.06864.pdf GitHub

Abstract:
Video Instance Segmentation (VIS) faces significant annotation challenges due to its dual requirements of pixel-level masks and temporal consistency labels. While recent unsupervised methods like VideoCutLER eliminate optical flow dependencies through synthetic data, they remain constrained by the synthetic-to-real domain gap. We present AutoQ-VIS, a novel unsupervised framework that bridges this gap through quality-guided self-training. Our approach establishes a closed-loop system between pseudo-label generation and automatic quality assessment, enabling progressive adaptation from synthetic to real videos. Experiments demonstrate state-of-the-art performance with 52.6 $\text{AP}_{50}$ on YouTubeVIS-2019 $\texttt{val}$ set, surpassing the previous state-of-the-art VideoCutLER by 4.4%, while requiring no human annotations. This demonstrates the viability of quality-aware self-training for unsupervised VIS. We will release the code at https://github.com/wcbup/AutoQ-VIS.

Paperid: 89, https://arxiv.org/pdf/2512.06854.pdf GitHub

Abstract:
Power is the primary design objective of large-scale integrated circuits (ICs), especially for complex modern processors (i.e., CPUs). Accurate CPU power evaluation requires designers to go through the whole time-consuming IC implementation process, easily taking months. At the early design stage (e.g., architecture-level), classical power models are notoriously inaccurate. Recently, ML-based architecture-level power models have been proposed to boost accuracy, but the data availability is a severe challenge. Currently, there is no open-source dataset for this important ML application. A typical dataset generation process involves correct CPU design implementation and repetitive execution of power simulation flows, requiring significant design expertise, engineering effort, and execution time. Even private in-house datasets often fail to reflect realistic CPU design scenarios. In this work, we propose ArchPower, the first open-source dataset for architecture-level processor power modeling. We go through complex and realistic design flows to collect the CPU architectural information as features and the ground-truth simulated power as labels. Our dataset includes 200 CPU data samples, collected from 25 different CPU configurations when executing 8 different workloads. There are more than 100 architectural features in each data sample, including both hardware and event parameters. The label of each sample provides fine-grained power information, including the total design power and the power for each of the 11 components. Each power value is further decomposed into four fine-grained power groups: combinational logic power, sequential logic power, memory power, and clock power. ArchPower is available at https://github.com/hkust-zhiyao/ArchPower.

Paperid: 90, https://arxiv.org/pdf/2512.06838.pdf GitHub

Abstract:
Cooperative perception is critical for autonomous driving, overcoming the inherent limitations of a single vehicle, such as occlusions and constrained fields-of-view. However, current approaches sharing dense Bird's-Eye-View (BEV) features are constrained by quadratically-scaling communication costs and the lack of flexibility and interpretability for precise alignment across asynchronous or disparate viewpoints. While emerging sparse query-based methods offer an alternative, they often suffer from inadequate geometric representations, suboptimal fusion strategies, and training instability. In this paper, we propose SparseCoop, a fully sparse cooperative perception framework for 3D detection and tracking that completely discards intermediate BEV representations. Our framework features a trio of innovations: a kinematic-grounded instance query that uses an explicit state vector with 3D geometry and velocity for precise spatio-temporal alignment; a coarse-to-fine aggregation module for robust fusion; and a cooperative instance denoising task to accelerate and stabilize training. Experiments on V2X-Seq and Griffin datasets show SparseCoop achieves state-of-the-art performance. Notably, it delivers this with superior computational efficiency, low transmission cost, and strong robustness to communication latency. Code is available at https://github.com/wang-jh18-SVM/SparseCoop.

Paperid: 91, https://arxiv.org/pdf/2512.06829.pdf GitHub GitHub

Abstract:
Vision-based tactile sensors (VBTS) face a fundamental trade-off in marker and markerless design on the tactile skin: opaque ink markers enable measurement of force and tangential displacement but completely occlude geometric features necessary for object and texture classification, while markerless skin preserves surface details but struggles in measuring tangential displacements effectively. Current practice to solve the above problem via UV lighting or virtual transfer using learning-based models introduces hardware complexity or computing burdens. This paper introduces MagicSkin, a novel tactile skin with translucent, tinted markers balancing the modes of marker and markerless for VBTS. It enables simultaneous tangential displacement tracking, force prediction, and surface detail preservation. This skin is easy to plug into GelSight-family sensors without requiring additional hardware or software tools. We comprehensively evaluate MagicSkin in downstream tasks. The translucent markers impressively enhance rather than degrade sensing performance compared with traditional markerless and inked marker design: it achieves best performance in object classification (99.17\%), texture classification (93.51\%), tangential displacement tracking (97\% point retention) and force prediction (66\% improvement in total force error). These experimental results demonstrate that translucent skin eliminates the traditional performance trade-off in marker or markerless modes, paving the way for multimodal tactile sensing essential in tactile robotics. See videos at this \href{https://zhuochenn.github.io/MagicSkin_project/}{link}.

Paperid: 92, https://arxiv.org/pdf/2512.06818.pdf GitHub

Abstract:
Primitive-based splatting methods like 3D Gaussian Splatting have revolutionized novel view synthesis with real-time rendering. However, their point-based representations remain incompatible with mesh-based pipelines that power AR/VR and game engines. We present MeshSplatting, a mesh-based reconstruction approach that jointly optimizes geometry and appearance through differentiable rendering. By enforcing connectivity via restricted Delaunay triangulation and refining surface consistency, MeshSplatting creates end-to-end smooth, visually high-quality meshes that render efficiently in real-time 3D engines. On Mip-NeRF360, it boosts PSNR by +0.69 dB over the current state-of-the-art MiLo for mesh-based novel view synthesis, while training 2x faster and using 2x less memory, bridging neural rendering and interactive 3D graphics for seamless real-time scene interaction. The project page is available at https://meshsplatting.github.io/.

Paperid: 93, https://arxiv.org/pdf/2512.06814.pdf GitHub

Abstract:
Multimodal classifiers function as opaque black box models. While several techniques exist to interpret their predictions, very few of them are as intuitive and accessible as natural language explanations (NLEs). To build trust, such explanations must faithfully capture the classifier's internal decision making behavior, a property known as faithfulness. In this paper, we propose CAuSE (Causal Abstraction under Simulated Explanations), a novel framework to generate faithful NLEs for any pretrained multimodal classifier. We demonstrate that CAuSE generalizes across datasets and models through extensive empirical evaluations. Theoretically, we show that CAuSE, trained via interchange intervention, forms a causal abstraction of the underlying classifier. We further validate this through a redesigned metric for measuring causal faithfulness in multimodal settings. CAuSE surpasses other methods on this metric, with qualitative analysis reinforcing its advantages. We perform detailed error analysis to pinpoint the failure cases of CAuSE. For replicability, we make the codes available at https://github.com/newcodevelop/CAuSE

Paperid: 94, https://arxiv.org/pdf/2512.06776.pdf GitHub

Abstract:
Large language models (LLMs) excel at generation but dominant autoregressive (AR) decoding is inherently sequential, creating a throughput bottleneck. Diffusion Language Models (DLMs)--especially block-wise variants--enable parallel generation and intra-block bidirectional reasoning, yet training large DLMs from scratch is costly and wastes the knowledge in mature AR checkpoints. Prior "adaptation" attempts either modify logits or randomly grow attention masks to full-sequence diffusion, or simply transplant AR weights into a block-diffusion recipe, leaving a fundamental mismatch between AR causality and block-wise bidirectionality unaddressed. We reframe adaptation as a intra-paradigm path from AR to Block-Diffusion by viewing AR as Block-Diffusion with blocksize=1. Concretely, we design the pathway of adaptation as follows: we use a context-causal attention mask (causal in context, bidirectional only within the active block), an efficient parallel adaptation procedure, an auxiliary AR loss to maximize data utilization and retain pretrained knowledge, and gradual increment of the generation block size. The recipe integrates cleanly with masked block-diffusion and maintains train-inference consistency. Built on these components, NBDiff-7B (Base and Instruct) could inherit the long-context modeling and reasoning capabilities, and achieve state-of-the-art performance among the 7B-class DLMs, delivering strong gains on general-knowledge, math, and code benchmarks over strong baselines. These results demonstrate that principled AR-to-block-diffusion adaptation is an effective and compute-efficient alternative to training DLMs from scratch. Codes: https://github.com/YuchuanTian/NBDiff.

Paperid: 95, https://arxiv.org/pdf/2512.06757.pdf GitHub

Abstract:
This paper introduces our solution, XM-ALIGN (Unified Cross-Modal Embedding Alignment Framework), proposed for the FAME challenge at ICASSP 2026. Our framework combines explicit and implicit alignment mechanisms, significantly improving cross-modal verification performance in both "heard" and "unheard" languages. By extracting feature embeddings from both face and voice encoders and jointly optimizing them using a shared classifier, we employ mean squared error (MSE) as the embedding alignment loss to ensure tight alignment between modalities. Additionally, data augmentation strategies are applied during model training to enhance generalization. Experimental results show that our approach demonstrates superior performance on the MAV-Celeb dataset. The code will be released at https://github.com/PunkMale/XM-ALIGN.

Paperid: 96, https://arxiv.org/pdf/2512.06752.pdf GitHub

Abstract:
Geometric Graph Neural Networks (GNNs) and Transformers have become state-of-the-art for learning from 3D protein structures. However, their reliance on message passing prevents them from capturing the hierarchical interactions that govern protein function, such as global domains and long-range allosteric regulation. In this work, we argue that the network architecture itself should mirror this biological hierarchy. We introduce Geometric Graph U-Nets, a new class of models that learn multi-scale representations by recursively coarsening and refining the protein graph. We prove that this hierarchical design can theoretically more expressive than standard Geometric GNNs. Empirically, on the task of protein fold classification, Geometric U-Nets substantially outperform invariant and equivariant baselines, demonstrating their ability to learn the global structural patterns that define protein folds. Our work provides a principled foundation for designing geometric deep learning architectures that can learn the multi-scale structure of biomolecules.

Paperid: 97, https://arxiv.org/pdf/2512.06750.pdf GitHub

Abstract:
Image quality assessment (IQA) and image restoration are fundamental problems in low-level vision. Although IQA and restoration are closely connected conceptually, most existing work treats them in isolation. Recent advances in unified multimodal understanding-generation models demonstrate promising results and indicate that stronger understanding can improve generative performance. This motivates a single model that unifies IQA and restoration and explicitly studies how IQA can guide restoration, a setting that remains largely underexplored yet highly valuable. In this paper, we propose UARE, to our knowledge the first Unified vision-language model for image quality Assessment, Restoration, and Enhancement. Built on pretrained unified understanding and generation models, we introduce a two-stage training framework. First, a progressive, easy-to-hard schedule expands from single-type distortions to higher-order mixed degradations, enabling UARE to handle multiple degradations. Second, we perform unified fine-tuning of quality understanding and restoration with interleaved text-image data, aligning IQA signals with restoration objectives. Through multi-task co-training, UARE leverages IQA to boost restoration and enhancement performance. Extensive experiments across IQA, restoration, and enhancement tasks demonstrate the effectiveness of UARE. The code and models will be available at https://github.com/lwq20020127/UARE.

Paperid: 98, https://arxiv.org/pdf/2512.06725.pdf GitHub

Abstract:
We present a novel approach to EEG decoding for non-invasive brain machine interfaces (BMIs), with a focus on motor-behavior classification. While conventional convolutional architectures such as EEGNet and DeepConvNet are effective in capturing local spatial patterns, they are markedly less suited for modeling long-range temporal dependencies and nonlinear dynamics. To address this limitation, we integrate an Echo State Network (ESN), a prominent paradigm in reservoir computing into the decoding pipeline. ESNs construct a high-dimensional, sparsely connected recurrent reservoir that excels at tracking temporal dynamics, thereby complementing the spatial representational power of CNNs. Evaluated on a skateboard-trick EEG dataset preprocessed via the PREP pipeline and implemented in MNE-Python, our ESNNet achieves 83.2% within-subject and 51.3% LOSO accuracies, surpassing widely used CNN-based baselines. Code is available at https://github.com/Yutiankunkun/Motion-Decoding-Using-Biosignals

Paperid: 99, https://arxiv.org/pdf/2512.06710.pdf GitHub

Abstract:
As large language models become components of larger agentic systems, evaluation reliability becomes critical: unreliable sub-agents introduce brittleness into downstream system behavior. Yet current evaluation practice, reporting a single accuracy number from a single run, obscures the variance underlying these results, making it impossible to distinguish genuine capability improvements from lucky sampling. We propose adopting Intraclass Correlation Coefficient (ICC), a metric from measurement science, to characterize this variance. ICC decomposes observed variance into between-query variance (task difficulty) and within-query variance (agent inconsistency), highlighting whether reported results reflect true capability or measurement noise. We evaluated on GAIA (Levels 1-3, measuring agentic capabilities across varying reasoning complexity) and FRAMES (measuring retrieval and factuality across multiple documents). We found that ICC varies dramatically with task structure, with reasoning and retrieval tasks (FRAMES) exhibit ICC=0.4955-0.7118 across models, and agentic tasks (GAIA) exhibiting ICC=0.304-0.774 across models. For sub-agent replacement decisions in agentic systems, accuracy improvements are only trustworthy if ICC also improves. We demonstrate that ICC converges by n=8-16 trials for structured tasks and n>=32 for complex reasoning, enabling practitioners to set evidence-based resampling budgets. We recommend reporting accuracy alongside ICC and within-query variance as standard practice, and propose updated Evaluation Cards capturing these metrics. By making evaluation stability visible, we aim to transform agentic benchmarking from opaque leaderboard competition to trustworthy experimental science. Our code is open-sourced at https://github.com/youdotcom-oss/stochastic-agent-evals.

Paperid: 100, https://arxiv.org/pdf/2512.06699.pdf GitHub

Abstract:
Modern machine learning training is increasingly bottlenecked by data I/O rather than compute. GPUs often sit idle at below 50% utilization waiting for data. This paper presents a machine learning approach to predict I/O performance and recommend optimal storage configurations for ML training pipelines. We collected 141 observations through systematic benchmarking across different storage backends (NVMe SSD, network-attached storage, in-memory filesystems), data formats, and access patterns, covering both low-level I/O operations and full training pipelines. After evaluating seven regression models and three classification approaches, XGBoost achieved the best performance with R-squared of 0.991, predicting I/O throughput within 11.8% error on average. Feature importance analysis revealed that throughput metrics and batch size are the primary performance drivers. This data-driven approach can reduce configuration time from days of trial-and-error to minutes of predictive recommendation. The methodology is reproducible and extensible to other resource management problems in ML systems. Code and data are available at https://github.com/knkarthik01/gpu_storage_ml_project

Paperid: 101, https://arxiv.org/pdf/2512.06694.pdf GitHub

Abstract:
Rapid expansion of social media platforms such as X (formerly Twitter), Facebook, and Reddit has enabled large-scale analysis of public perceptions on diverse topics, including social issues, politics, natural disasters, and consumer sentiment. Topic modeling is a widely used approach for uncovering latent themes in text data, typically framed as an unsupervised classification task. However, traditional models, originally designed for longer and more formal documents, struggle with short social media posts due to limited co-occurrence statistics, fragmented semantics, inconsistent spelling, and informal language. To address these challenges, we propose a new method, TopiCLEAR: Topic extraction by CLustering Embeddings with Adaptive dimensional Reduction. Specifically, each text is embedded using Sentence-BERT (SBERT) and provisionally clustered using Gaussian Mixture Models (GMM). The clusters are then refined iteratively using a supervised projection based on linear discriminant analysis, followed by GMM-based clustering until convergence. Notably, our method operates directly on raw text, eliminating the need for preprocessing steps such as stop word removal. We evaluate our approach on four diverse datasets, 20News, AgNewsTitle, Reddit, and TweetTopic, each containing human-labeled topic information. Compared with seven baseline methods, including a recent SBERT-based method and a zero-shot generative AI method, our approach achieves the highest similarity to human-annotated topics, with significant improvements for both social media posts and online news articles. Additionally, qualitative analysis shows that our method produces more interpretable topics, highlighting its potential for applications in social media data and web content analytics.

Paperid: 102, https://arxiv.org/pdf/2512.06689.pdf GitHub

Abstract:
Speech Enhancement (SE) and Speech Separation (SS) have traditionally been treated as distinct tasks in speech processing. However, real-world audio often involves both background noise and overlapping speakers, motivating the need for a unified solution. While recent approaches have attempted to integrate SE and SS within multi-stage architectures, these approaches typically involve complex, parameter-heavy models and rely on supervised training, limiting scalability and generalization. In this work, we propose UniVoiceLite, a lightweight and unsupervised audio-visual framework that unifies SE and SS within a single model. UniVoiceLite leverages lip motion and facial identity cues to guide speech extraction and employs Wasserstein distance regularization to stabilize the latent space without requiring paired noisy-clean data. Experimental results demonstrate that UniVoiceLite achieves strong performance in both noisy and multi-speaker scenarios, combining efficiency with robust generalization. The source code is available at https://github.com/jisoo-o/UniVoiceLite.

Paperid: 103, https://arxiv.org/pdf/2512.06673.pdf GitHub

Abstract:
Spatio-temporal grounding and reasoning aims to locate the temporal segment and spatial region of an event in a video given a user query, while also reasoning about semantics such as causality, temporal order, and action relationships. To achieve this, current MLLMs primarily treats bounding boxes as text tokens and generates them autoregressively. However, such autoregressive spatial decoding leads to very-long output sequences, causing spatial errors to accumulated over time and the localization results to progressively drift across a video. To address this, we present a Detector-Empowered Video LLM, short for DEViL, which couples a Video LLM with an open-vocabulary detector (OVD). Specifically, the MLLM and detector are connected via a reference-semantic token (RST) that distills the user query into a rich semantic representation. Unlike tokens that merely serve as spatial prompts or segmentor switches, the RST functions as both a control signal and a replacement for the OVD's text embedding, enabling end-to-end learning of both referential understanding and spatial localization. Furthermore, we propose a tube-mined temporal regularization (TTReg) within OVD, which drives the OVD to generate temporally-consistent queries for target objects, thereby ensuring effective temporal association. Experiments demonstrate that DEViL achieves strong performance across various fine-grained video understanding tasks, particularly STVG and GroundedVQA. Code will be released on https://github.com/gaostar123/DeViL.

Paperid: 104, https://arxiv.org/pdf/2512.06666.pdf GitHub

Abstract:
Time series classification faces a fundamental trade-off between accuracy and computational efficiency. While comprehensive ensembles like HIVE-COTE 2.0 achieve state-of-the-art accuracy, their 340-hour training time on the UCR benchmark renders them impractical for large-scale datasets. We investigate whether targeted combinations of two efficient algorithms from complementary paradigms can capture ensemble benefits while maintaining computational feasibility. Combining Hydra (competing convolutional kernels) and Quant (hierarchical interval quantiles) across six ensemble configurations, we evaluate performance on 10 large-scale MONSTER datasets (7,898 to 1,168,774 training instances). Our strongest configuration improves mean accuracy from 0.829 to 0.836, succeeding on 7 of 10 datasets. However, prediction-combination ensembles capture only 11% of theoretical oracle potential, revealing a substantial meta-learning optimization gap. Feature-concatenation approaches exceeded oracle bounds by learning novel decision boundaries, while prediction-level complementarity shows moderate correlation with ensemble gains. The central finding: the challenge has shifted from ensuring algorithms are different to learning how to combine them effectively. Current meta-learning strategies struggle to exploit the complementarity that oracle analysis confirms exists. Improved combination strategies could potentially double or triple ensemble gains across diverse time series classification applications.

Paperid: 105, https://arxiv.org/pdf/2512.06612.pdf GitHub

Abstract:
Gene expression estimation from pathology images has the potential to reduce the RNA sequencing cost. Point-wise loss functions have been widely used to minimize the discrepancy between predicted and absolute gene expression values. However, due to the complexity of the sequencing techniques and intrinsic variability across cells, the observed gene expression contains stochastic noise and batch effects, and estimating the absolute expression values accurately remains a significant challenge. To mitigate this, we propose a novel objective of learning relative expression patterns rather than absolute levels. We assume that the relative expression levels of genes exhibit consistent patterns across independent experiments, even when absolute expression values are affected by batch effects and stochastic noise in tissue samples. Based on the assumption, we model the relation and propose a novel loss function called STRank that is robust to noise and batch effects. Experiments using synthetic datasets and real datasets demonstrate the effectiveness of the proposed method. The code is available at https://github.com/naivete5656/STRank.

Paperid: 106, https://arxiv.org/pdf/2512.06609.pdf GitHub

Abstract:
Vector quantized variational autoencoder (VQ-VAE) is a discrete auto-encoder that compresses images into discrete tokens. It is difficult to train due to discretization. In this paper, we propose a simple yet effective technique, dubbed Gaussian Quant (GQ), that converts a Gaussian VAE with certain constraint into a VQ-VAE without training. GQ generates random Gaussian noise as a codebook and finds the closest noise to the posterior mean. Theoretically, we prove that when the logarithm of the codebook size exceeds the bits-back coding rate of the Gaussian VAE, a small quantization error is guaranteed. Practically, we propose a heuristic to train Gaussian VAE for effective GQ, named target divergence constraint (TDC). Empirically, we show that GQ outperforms previous VQ-VAEs, such as VQGAN, FSQ, LFQ, and BSQ, on both UNet and ViT architectures. Furthermore, TDC also improves upon previous Gaussian VAE discretization methods, such as TokenBridge. The source code is provided in https://github.com/tongdaxu/VQ-VAE-from-Gaussian-VAE.

Paperid: 107, https://arxiv.org/pdf/2512.06591.pdf GitHub

Abstract:
Explainable AI (XAI) presents useful tools to facilitate transparency and trustworthiness in machine learning systems. However, current evaluations of system explainability often rely heavily on subjective user surveys, which may not adequately capture the effectiveness of explanations. This paper critiques the overreliance on user satisfaction metrics and explores whether these can differentiate between meaningful (actionable) and vacuous (placebic) explanations. In experiments involving optimal Social Security filing age selection tasks, participants used one of three protocols: no explanations, placebic explanations, and actionable explanations. Participants who received actionable explanations significantly outperformed the other groups in objective measures of their mental model, but users rated placebic and actionable explanations as equally satisfying. This suggests that subjective surveys alone fail to capture whether explanations truly support users in building useful domain understanding. We propose that future evaluations of agent explanation capabilities should integrate objective task performance metrics alongside subjective assessments to more accurately measure explanation quality. The code for this study can be found at https://github.com/Shymkis/social-security-explainer.

Paperid: 108, https://arxiv.org/pdf/2512.06589.pdf GitHub

Abstract:
Recent advances in multi-modal large language models (MLLMs) have enabled unified perception-reasoning capabilities, yet these systems remain highly vulnerable to jailbreak attacks that bypass safety alignment and induce harmful behaviors. Existing benchmarks such as JailBreakV-28K, MM-SafetyBench, and HADES provide valuable insights into multi-modal vulnerabilities, but they typically focus on limited attack scenarios, lack standardized defense evaluation, and offer no unified, reproducible toolbox. To address these gaps, we introduce OmniSafeBench-MM, which is a comprehensive toolbox for multi-modal jailbreak attack-defense evaluation. OmniSafeBench-MM integrates 13 representative attack methods, 15 defense strategies, and a diverse dataset spanning 9 major risk domains and 50 fine-grained categories, structured across consultative, imperative, and declarative inquiry types to reflect realistic user intentions. Beyond data coverage, it establishes a three-dimensional evaluation protocol measuring (1) harmfulness, distinguished by a granular, multi-level scale ranging from low-impact individual harm to catastrophic societal threats, (2) intent alignment between responses and queries, and (3) response detail level, enabling nuanced safety-utility analysis. We conduct extensive experiments on 10 open-source and 8 closed-source MLLMs to reveal their vulnerability to multi-modal jailbreak. By unifying data, methodology, and evaluation into an open-source, reproducible platform, OmniSafeBench-MM provides a standardized foundation for future research. The code is released at https://github.com/jiaxiaojunQAQ/OmniSafeBench-MM.

Paperid: 109, https://arxiv.org/pdf/2512.06581.pdf GitHub

Abstract:
Large vision-language models struggle with medical video understanding, where spatial precision, temporal reasoning, and clinical semantics are critical. To address this, we first introduce \textbf{MedVidBench}, a large-scale benchmark of 531,850 video-instruction pairs across 8 medical sources spanning video, segment, and frame-level tasks, curated through a rigorous quality assurance pipeline with expert-guided prompting and dual-model validation. While supervised fine-tuning on MedVidBench yields noticeable gains, standard Reinforcement Learning (RL) fails due to imbalanced reward scales across datasets, which destabilizes optimization and leads to training collapse. To overcome this, we introduce \textbf{MedGRPO}, a novel RL framework for balanced multi-dataset training with two key innovations: (1) \emph{cross-dataset reward normalization} that maps each dataset's median performance to a common reward value, ensuring fair optimization regardless of difficulty, and (2) a \emph{medical LLM judge} that evaluates caption quality on five clinical dimensions through comparative similarity scoring. Supervised fine-tuning Qwen2.5-VL-7B on MedVidBench substantially outperforms GPT-4.1 and Gemini-2.5-Flash across all tasks, demonstrating MedVidBench's efficacy, while our MedGRPO framework further improves upon the SFT baseline across grounding and captioning tasks. Our work establishes a foundational benchmark and robust training methodology for advancing vision-language models in medical domains. Our project website is available at https://yuhaosu.github.io/MedGRPO/.

Paperid: 110, https://arxiv.org/pdf/2512.06562.pdf GitHub

Abstract:
Recent advances in 3D-aware generative models have enabled high-fidelity image synthesis of human identities. However, this progress raises urgent questions around user consent and the ability to remove specific individuals from a model's output space. We address this by introducing SUGAR, a framework for scalable generative unlearning that enables the removal of many identities (simultaneously or sequentially) without retraining the entire model. Rather than projecting unwanted identities to unrealistic outputs or relying on static template faces, SUGAR learns a personalized surrogate latent for each identity, diverting reconstructions to visually coherent alternatives while preserving the model's quality and diversity. We further introduce a continual utility preservation objective that guards against degradation as more identities are forgotten. SUGAR achieves state-of-the-art performance in removing up to 200 identities, while delivering up to a 700% improvement in retention utility compared to existing baselines. Our code is publicly available at https://github.com/judydnguyen/SUGAR-Generative-Unlearn.

Paperid: 111, https://arxiv.org/pdf/2512.06547.pdf GitHub

Abstract:
Decoupled loss has been a successful reinforcement learning (RL) algorithm to deal with the high data staleness under the asynchronous RL setting. Decoupled loss improves coupled-loss style of algorithms' (e.g., PPO, GRPO) learning stability by introducing a proximal policy to decouple the off-policy corrections (importance weight) from the controlling policy updates (trust region). However, the proximal policy requires an extra forward pass through the network at each training step, creating a computational bottleneck for large language models. We observe that since the proximal policy only serves as a trust region anchor between the behavior and target policies, we can approximate it through simple interpolation without explicit computation. We call this approach A-3PO (APproximated Proximal Policy Optimization). A-3PO eliminates this overhead, reducing training time by 18% while maintaining comparable performance. Code & off-the-shelf example are available at: https://github.com/inclusionAI/AReaL/blob/main/docs/algorithms/prox_approx.md

Paperid: 112, https://arxiv.org/pdf/2512.06443.pdf GitHub

Abstract:
Large language models (LLMs) are increasingly deployed on edge devices. To meet strict resource constraints, real-world deployment has pushed LLM quantization from 8-bit to 4-bit, 2-bit, and now 1.58-bit. Combined with lookup table (LUT)-based inference, CPUs run these ultra-low-bit LLMs even faster than NPUs, opening new opportunities for ubiquitous on-device intelligence. However, this paper identifies that LUT-based inference underutilizes memory bandwidth during parallel inference, which is required for prefilling, test-time scaling, and other multi-token scenarios. The root cause is the scalar LUT paradigm, which performs repetitive and non-contiguous memory accesses for each token. To solve the issue, we propose vector LUT, a new lookup paradigm that constructs a unified LUT across parallel tokens, and performs a single $1 \rightarrow N$ lookup per index. To realize it efficiently, we further introduce (1) Vector LUT-Centric Tensor Layout, and (2) Cache-Aware Streamed Lookup techniques. Evaluations on 5 edge devices across 3 LLMs show that Vec-LUT outperforms state-of-the-art baselines by up to $4.2\times$. Our implementation is integrated into llama.cpp. The code is available at https://github.com/Cipherxzc/vlut.cpp.

Paperid: 113, https://arxiv.org/pdf/2512.06438.pdf GitHub

Abstract:
The generation of high-fidelity, animatable 3D human avatars remains a core challenge in computer graphics and vision, with applications in VR, telepresence, and entertainment. Existing approaches based on implicit representations like NeRFs suffer from slow rendering and dynamic inconsistencies, while 3D Gaussian Splatting (3DGS) methods are typically limited to static head generation, lacking dynamic control. We bridge this gap by introducing AGORA, a novel framework that extends 3DGS within a generative adversarial network to produce animatable avatars. Our key contribution is a lightweight, FLAME-conditioned deformation branch that predicts per-Gaussian residuals, enabling identity-preserving, fine-grained expression control while allowing real-time inference. Expression fidelity is enforced via a dual-discriminator training scheme leveraging synthetic renderings of the parametric mesh. AGORA generates avatars that are not only visually realistic but also precisely controllable. Quantitatively, we outperform state-of-the-art NeRF-based methods on expression accuracy while rendering at 250+ FPS on a single GPU, and, notably, at $\sim$9 FPS under CPU-only inference - representing, to our knowledge, the first demonstration of practical CPU-only animatable 3DGS avatar synthesis. This work represents a significant step toward practical, high-performance digital humans. Project website: https://ramazan793.github.io/AGORA/

Paperid: 114, https://arxiv.org/pdf/2512.06424.pdf GitHub GitHub

Abstract:
While generative models have excelled at creating static 3D content, the pursuit of systems that understand how objects move and respond to interactions remains a fundamental challenge. Current methods for articulated motion lie at a crossroads: they are either physically consistent but too slow for real-time use, or generative but violate basic kinematic constraints. We present DragMesh, a robust framework for real-time interactive 3D articulation built around a lightweight motion generation core. Our core contribution is a novel decoupled kinematic reasoning and motion generation framework. First, we infer the latent joint parameters by decoupling semantic intent reasoning (which determines the joint type) from geometric regression (which determines the axis and origin using our Kinematics Prediction Network (KPP-Net)). Second, to leverage the compact, continuous, and singularity-free properties of dual quaternions for representing rigid body motion, we develop a novel Dual Quaternion VAE (DQ-VAE). This DQ-VAE receives these predicted priors, along with the original user drag, to generate a complete, plausible motion trajectory. To ensure strict adherence to kinematics, we inject the joint priors at every layer of the DQ-VAE's non-autoregressive Transformer decoder using FiLM (Feature-wise Linear Modulation) conditioning. This persistent, multi-scale guidance is complemented by a numerically-stable cross-product loss to guarantee axis alignment. This decoupled design allows DragMesh to achieve real-time performance and enables plausible, generative articulation on novel objects without retraining, offering a practical step toward generative 3D intelligence. Code: https://github.com/AIGeeksGroup/DragMesh. Website: https://aigeeksgroup.github.io/DragMesh.

Paperid: 115, https://arxiv.org/pdf/2512.06422.pdf GitHub

Abstract:
Convolutional neural networks (CNNs) can automatically learn data patterns to express face images for facial expression recognition (FER). However, they may ignore effect of facial segmentation of FER. In this paper, we propose a perception CNN for FER as well as PCNN. Firstly, PCNN can use five parallel networks to simultaneously learn local facial features based on eyes, cheeks and mouth to realize the sensitive capture of the subtle changes in FER. Secondly, we utilize a multi-domain interaction mechanism to register and fuse between local sense organ features and global facial structural features to better express face images for FER. Finally, we design a two-phase loss function to restrict accuracy of obtained sense information and reconstructed face images to guarantee performance of obtained PCNN in FER. Experimental results show that our PCNN achieves superior results on several lab and real-world FER benchmarks: CK+, JAFFE, FER2013, FERPlus, RAF-DB and Occlusion and Pose Variant Dataset. Its code is available at https://github.com/hellloxiaotian/PCNN.

Paperid: 116, https://arxiv.org/pdf/2512.06411.pdf GitHub

Abstract:
KyFrog is a conservative Learning-with-Errors (LWE) key-encapsulation mechanism designed to explore an alternative operating point compared to schemes with relatively small public keys and ciphertexts. KyFrog uses a larger dimension (n = 1024) and a small prime modulus q = 1103, together with narrow error distributions with standard deviations sigma_s = sigma_e = 1.4, to target approximately 2^325 classical and quantum security against state-of-the-art lattice attacks under standard cost models, as estimated using the Lattice Estimator. The price paid for this security margin is an extremely large KEM ciphertext (about 0.5 MiB), while public and secret keys remain in the same ballpark as ML-KEM. We describe the design rationale, parameter search methodology, and implementation details of KyFrog, and we compare its asymptotic security and concrete parameter sizes with the ML-KEM standard. All code and data for this work are released as free and open-source software, with the full C++23 implementation and experimental scripts available at: https://github.com/victormeloasm/kyfrog

Paperid: 117, https://arxiv.org/pdf/2512.06406.pdf GitHub

Abstract:
Large language models(LLMs) are increasingly expanding their real-world applications across domains, e.g., question answering, autonomous driving, and automatic software development. Despite this achievement, LLMs, as data-driven systems, often make incorrect predictions, which can lead to potential losses in safety-critical scenarios. To address this issue and measure the confidence of model outputs, multiple uncertainty quantification(UQ) criteria have been proposed. However, even though important, there are limited tools to integrate these methods, hindering the practical usage of UQ methods and future research in this domain. To bridge this gap, in this paper, we introduce UncertaintyZoo, a unified toolkit that integrates 29 uncertainty quantification methods, covering five major categories under a standardized interface. Using UncertaintyZoo, we evaluate the usefulness of existing uncertainty quantification methods under the code vulnerability detection task on CodeBERT and ChatGLM3 models. The results demonstrate that UncertaintyZoo effectively reveals prediction uncertainty. The tool with a demonstration video is available on the project site https://github.com/Paddingbuta/UncertaintyZoo.

Paperid: 118, https://arxiv.org/pdf/2512.06379.pdf GitHub

Abstract:
Recently, online learning is very popular, especially under the global epidemic of COVID-19. Besides knowledge distribution, emotion interaction is also very important. It can be obtained by employing Facial Expression Recognition (FER). Since the FER accuracy is substantial in assisting teachers to acquire the emotional situation, the project explores a series of FER methods and finds that few works engage in exploiting the orthogonality of convolutional matrix. Therefore, it enforces orthogonality on kernels by a regularizer, which extracts features with more diversity and expressiveness, and delivers OCFER-Net. Experiments are carried out on FER-2013, which is a challenging dataset. Results show superior performance over baselines by 1.087. The code of the research project is publicly available on https://github.com/YeeHoran/OCFERNet.

Paperid: 119, https://arxiv.org/pdf/2512.06377.pdf GitHub

Abstract:
Current FER (Facial Expression Recognition) dataset is mostly labeled by emotion categories, such as happy, angry, sad, fear, disgust, surprise, and neutral which are limited in expressiveness. However, future affective computing requires more comprehensive and precise emotion metrics which could be measured by VAD(Valence-Arousal-Dominance) multidimension parameters. To address this, AffectNet has tried to add VA (Valence and Arousal) information, but still lacks D(Dominance). Thus, the research introduces VAD annotation on FER2013 dataset, takes the initiative to label D(Dominance) dimension. Then, to further improve network capacity, it enforces orthogonalized convolution on it, which extracts more diverse and expressive features and will finally increase the prediction accuracy. Experiment results show that D dimension could be measured but is difficult to obtain compared with V and A dimension no matter in manual annotation or regression network prediction. Secondly, the ablation test by introducing orthogonal convolution verifies that better VAD prediction could be obtained in the configuration of orthogonal convolution. Therefore, the research provides an initiative labelling for D dimension on FER dataset, and proposes a better prediction network for VAD prediction through orthogonal convolution. The newly built VAD annotated FER2013 dataset could act as a benchmark to measure VAD multidimensional emotions, while the orthogonalized regression network based on ResNet could act as the facial expression recognition baseline for VAD emotion prediction. The newly labeled dataset and implementation code is publicly available on https://github.com/YeeHoran/VAD-Net .

Paperid: 120, https://arxiv.org/pdf/2512.06373.pdf GitHub

Abstract:
Tool-integrated visual reasoning (TiVR) has demonstrated great potential in enhancing multimodal problem-solving. However, existing TiVR paradigms mainly focus on integrating various visual tools through reinforcement learning, while neglecting to design effective response mechanisms for handling unreliable or erroneous tool outputs. This limitation is particularly pronounced in referring and grounding tasks, where inaccurate detection tool predictions often mislead TiVR models into generating hallucinated reasoning. To address this issue, we propose the VG-Refiner, the first framework aiming at the tool-refined referring grounded reasoning. Technically, we introduce a two-stage think-rethink mechanism that enables the model to explicitly analyze and respond to tool feedback, along with a refinement reward that encourages effective correction in response to poor tool results. In addition, we propose two new metrics and establish fair evaluation protocols to systematically measure the refinement ability of current models. We adopt a small amount of task-specific data to enhance the refinement capability of VG-Refiner, achieving a significant improvement in accuracy and correction ability on referring and reasoning grounding benchmarks while preserving the general capabilities of the pretrained model.

Paperid: 121, https://arxiv.org/pdf/2512.06353.pdf GitHub

Abstract:
Diffusion Transformers (DiTs) have emerged as a highly scalable and effective backbone for image generation, outperforming U-Net architectures in both scalability and performance. However, their real-world deployment remains challenging due to high computational and memory demands. Mixed-Precision Quantization (MPQ), designed to push the limits of quantization, has demonstrated remarkable success in advancing U-Net quantization to sub-4bit settings while significantly reducing computational and memory overhead. Nevertheless, its application to DiT architectures remains limited and underexplored. In this work, we propose TreeQ, a unified framework addressing key challenges in DiT quantization. First, to tackle inefficient search and proxy misalignment, we introduce Tree Structured Search (TSS). This DiT-specific approach leverages the architecture's linear properties to traverse the solution space in O(n) time while improving objective accuracy through comparison-based pruning. Second, to unify optimization objectives, we propose Environmental Noise Guidance (ENG), which aligns Post-Training Quantization (PTQ) and Quantization-Aware Training (QAT) configurations using a single hyperparameter. Third, to mitigate information bottlenecks in ultra-low-bit regimes, we design the General Monarch Branch (GMB). This structured sparse branch prevents irreversible information loss, enabling finer detail generation. Through extensive experiments, our TreeQ framework demonstrates state-of-the-art performance on DiT-XL/2 under W3A3 and W4A4 PTQ/PEFT settings. Notably, our work is the first to achieve near-lossless 4-bit PTQ performance on DiT models. The code and models will be available at https://github.com/racoonykc/TreeQ

Paperid: 122, https://arxiv.org/pdf/2512.06330.pdf GitHub

Abstract:
Pansharpening fuses a high-resolution PAN image with a low-resolution multispectral (LRMS) image to produce an HRMS image. A key difficulty is that jointly processing PAN and MS often entangles spatial detail with spectral fidelity. We propose S2WMamba, which explicitly disentangles frequency information and then performs lightweight cross-modal interaction. Concretely, a 2D Haar DWT is applied to PAN to localize spatial edges and textures, while a channel-wise 1D Haar DWT treats each pixel's spectrum as a 1D signal to separate low/high-frequency components and limit spectral distortion. The resulting Spectral branch injects wavelet-extracted spatial details into MS features, and the Spatial branch refines PAN features using spectra from the 1D pyramid; the two branches exchange information through Mamba-based cross-modulation that models long-range dependencies with linear complexity. A multi-scale dynamic gate (multiplicative + additive) then adaptively fuses branch outputs.On WV3, GF2, and QB, S2WMamba matches or surpasses recent strong baselines (FusionMamba, CANNet, U2Net, ARConv), improving PSNR by up to 0.23 dB and reaching HQNR 0.956 on full-resolution WV3. Ablations justify the choice of 2D/1D DWT placement, parallel dual branches, and the fusion gate. Our code is available at https://github.com/KagUYa66/S2WMamba.

Paperid: 123, https://arxiv.org/pdf/2512.06293.pdf GitHub

Abstract:
Urban transit agencies increasingly turn to social media to monitor emerging service risks such as crowding, delays, and safety incidents, yet the signals of concern are sparse, short, and easily drowned by routine chatter. We address this challenge by jointly modeling linguistic interactions and user influence. First, we construct an influence-weighted keyword co-occurrence graph from cleaned posts so that socially impactful posts contributes proportionally to the underlying evidence. The core of our framework is a Poisson Deconvolution Factorization (PDF) that decomposes this graph into a low-rank topical structure and topic-localized residual interactions, producing an interpretable topic--keyword basis together with topic importance scores. A decorrelation regularizer \emph{promotes} distinct topics, and a lightweight optimization procedure ensures stable convergence under nonnegativity and normalization constraints. Finally, the number of topics is selected through a coherence-driven sweep that evaluates the quality and distinctness of the learned topics. On large-scale social streams, the proposed model achieves state-of-the-art topic coherence and strong diversity compared with leading baselines. The code and dataset are publicly available at https://github.com/pangjunbiao/Topic-Modeling_ITS.git

Paperid: 124, https://arxiv.org/pdf/2512.06281.pdf GitHub

Abstract:
Multimodal Large Language Models (MLLMs) have demonstrated remarkable proficiency in multimodal tasks. Despite their impressive performance, MLLMs suffer from the modality imbalance issue, where visual information is often underutilized compared to textual representations in deeper layers, leading to degraded visual performance or hallucinations. This issue stems from the predominant reliance on next-text-token-prediction during training, which fails to provide direct visual supervisory signals, resulting in progressive homogenization of visual representations throughout the layers. To this end, we propose Latent Visual Reconstruction (LaVer), a novel training framework that facilitates MLLMs in learning more discriminative visual representations via masked image modeling in the joint latent semantic space of LLM. Our method offers direct visual activation to MLLMs, which exhibit increased visual attention allocation, indicating enhanced utilization of visual information. Extensive experiments across diverse benchmarks prove the superiority of our approach in various scenarios, especially those requiring dense visual capabilities. Code of LaVer is available at https://github.com/Fir-lat/LaVer.

Paperid: 125, https://arxiv.org/pdf/2512.06258.pdf GitHub

Abstract:
We reveal a critical yet underexplored flaw in Large Vision-Language Models (LVLMs): even when these models know the correct answer, they frequently arrive there through incorrect reasoning paths. The core issue is not a lack of knowledge, but a path selection bias within the vast reasoning search space. Although LVLMs are often capable of sampling correct solution trajectories, they disproportionately favor unstable or logically inconsistent ones, leading to erratic and unreliable outcomes. The substantial disparity between Pass@K (with large K) and Pass@1 across numerous models provides compelling evidence that such failures primarily stem from misreasoning rather than ignorance. To systematically investigate and address this issue, we propose PSO (Path-Select Optimization), a two-stage post-training framework designed to enhance both the reasoning performance and stability of existing LVLMs. In the first stage, we employ Group Relative Policy Optimization (GRPO) with template and answer-based rewards to cultivate structured, step-by-step reasoning. In the second stage, we conduct online preference optimization, where the model samples reasoning paths from GRPO-generated data, self-evaluates them, and aligns itself toward the preferred trajectories. Incorrect or suboptimal paths are concurrently stored in a Negative Replay Memory (NRM) as hard negatives, which are periodically revisited to prevent the model from repeating prior mistakes and to facilitate continual reasoning refinement. Extensive experiments show that PSO effectively prunes invalid reasoning paths, substantially enhances reasoning accuracy (with 7.4% improvements on average), and yields more stable and consistent chains of thought. Our code will be available at https://github.com/aiming-lab/PSO.

Paperid: 126, https://arxiv.org/pdf/2512.06210.pdf GitHub

Abstract:
Predictions of fatalities from violent conflict on the PRIO-GRID-month (pgm) level are characterized by high levels of uncertainty, limiting their usefulness in practical applications. We discuss the two main sources of uncertainty for this prediction task, the nature of violent conflict and data limitations, embedding this in the wider literature on uncertainty quantification in machine learning. We develop a strategy to quantify uncertainty in conflict forecasting, shifting from traditional point predictions to full predictive distributions. Our approach compares and combines multiple tree-based classifiers and distributional regressors in a custom auto-ML setup, estimating distributions for each pgm individually. We also test the integration of regional models in spatial ensembles as a potential avenue to reduce uncertainty. The models are able to consistently outperform a suite of benchmarks derived from conflict history in predictions up to one year in advance, with performance driven by regions where conflict was observed. With our evaluation, we emphasize the need to understand how a metric behaves for a given prediction problem, in our case characterized by extremely high zero-inflatedness. While not resulting in better predictions, the integration of smaller models does not decrease performance for this prediction task, opening avenues to integrate data sources with less spatial coverage in the future.

Paperid: 127, https://arxiv.org/pdf/2512.06206.pdf GitHub

Abstract:
We present the design and results of the MICCAI Federated Tumor Segmentation (FeTS) Challenge 2024, which focuses on federated learning (FL) for glioma sub-region segmentation in multi-parametric MRI and evaluates new weight aggregation methods aimed at improving robustness and efficiency. Six participating teams were evaluated using a standardized FL setup and a multi-institutional dataset derived from the BraTS glioma benchmark, consisting of 1,251 training cases, 219 validation cases, and 570 hidden test cases with segmentations for enhancing tumor (ET), tumor core (TC), and whole tumor (WT). Teams were ranked using a cumulative scoring system that considered both segmentation performance, measured by Dice Similarity Coefficient (DSC) and the 95th percentile Hausdorff Distance (HD95), and communication efficiency assessed through the convergence score. A PID-controller-based method achieved the top overall ranking, obtaining mean DSC values of 0.733, 0.761, and 0.751 for ET, TC, and WT, respectively, with corresponding HD95 values of 33.922 mm, 33.623 mm, and 32.309 mm, while also demonstrating the highest communication efficiency with a convergence score of 0.764. These findings advance the state of federated learning for medical imaging, surpassing top-performing methods from previous challenge iterations and highlighting PID controllers as effective mechanisms for stabilizing and optimizing weight aggregation in FL. The challenge code is available at https://github.com/FeTS-AI/Challenge.

Paperid: 128, https://arxiv.org/pdf/2512.06155.pdf GitHub

Abstract:
Security research is fundamentally a problem of resource constraint and consequent prioritization. There is simply too much attack surface and too little time and energy to spend analyzing it all. The most effective security researchers are often those who are most skilled at intuitively deciding which part of an expansive attack surface to investigate. We demonstrate that this problem of selecting the most promising option from among many possibilities can be reframed as an information retrieval problem, and solved using document ranking techniques with LLMs performing the heavy lifting as general-purpose rankers. We present SiftRank, a ranking algorithm achieving O(n) complexity through three key mechanisms: listwise ranking using an LLM to order documents in small batches of approximately 10 items at a time; inflection-based convergence detection that adaptively terminates ranking when score distributions have stabilized; and iterative refinement that progressively focuses ranking effort on the most relevant documents. Unlike existing reranking approaches that require a separate first-stage retrieval step to narrow datasets to approximately 100 candidates, SiftRank operates directly on thousands of items, with each document evaluated across multiple randomized batches to mitigate inconsistent judgments by an LLM. We demonstrate practical effectiveness on N-day vulnerability analysis, successfully identifying a vulnerability-fixing function among 2,197 changed functions in a stripped binary firmware patch within 99 seconds at an inference cost of $0.82. Our approach enables scalable security prioritization for problems that are generally constrained by manual analysis, requiring only standard LLM API access without specialized infrastructure, embedding, or domain-specific fine-tuning. An open-source implementation of SiftRank may be found at https://github.com/noperator/siftrank.

Paperid: 129, https://arxiv.org/pdf/2512.06097.pdf GitHub

Abstract:
General-purpose large language models (LLMs) have demonstrated remarkable generative and reasoning capabilities but remain limited in healthcare and caregiving applications due to two key deficiencies: factual unreliability and a lack of empathetic communication. These shortcomings pose significant risks in sensitive contexts where users, particularly non-professionals and caregivers, seek medically relevant guidance or emotional reassurance. To address these challenges, we introduce a Direct Preference Optimization (DPO)-based alignment framework designed to improve factual correctness, semantic coherence, and human-centric qualities such as empathy, politeness, and simplicity in caregiver-patient dialogues. Our approach fine-tunes domain-adapted LLMs using pairwise preference data, where preferred responses reflect supportive and accessible communication styles while rejected ones represent prescriptive or overly technical tones. This direct optimization method aligns model outputs with human preferences more efficiently than traditional reinforcement-learning-based alignment. Empirical evaluations across multiple open and proprietary LLMs show that our DPO-tuned models achieve higher semantic alignment, improved factual accuracy, and stronger human-centric evaluation scores compared to baseline and commercial alternatives such as Google medical dialogue systems. These improvements demonstrate that preference-based alignment offers a scalable and transparent pathway toward developing trustworthy, empathetic, and clinically informed AI assistants for caregiver and healthcare communication. Our open-source code is available at: https://github.com/LeonG19/Empathy-by-Design

Paperid: 130, https://arxiv.org/pdf/2512.06065.pdf GitHub

Abstract:
We study instruction-guided editing of egocentric videos for interactive AR applications. While recent AI video editors perform well on third-person footage, egocentric views present unique challenges - including rapid egomotion and frequent hand-object interactions - that create a significant domain gap. Moreover, existing offline editing pipelines suffer from high latency, limiting real-time interaction. To address these issues, we present a complete ecosystem for egocentric video editing. First, we construct EgoEditData, a carefully designed and manually curated dataset specifically designed for egocentric editing scenarios, featuring rich hand-object interactions, while explicitly preserving hands. Second, we develop EgoEdit, an instruction-following egocentric video editor that supports real-time streaming inference on a single GPU. Finally, we introduce EgoEditBench, an evaluation suite targeting instruction faithfulness, hand and interaction preservation, and temporal stability under egomotion. Across both egocentric and general editing tasks, EgoEdit produces temporally stable, instruction-faithful results with interactive latency. It achieves clear gains on egocentric editing benchmarks-where existing methods struggle-while maintaining performance comparable to the strongest baselines on general editing tasks. EgoEditData and EgoEditBench will be made public for the research community. See our website at https://snap-research.github.io/EgoEdit

Paperid: 131, https://arxiv.org/pdf/2512.06020.pdf GitHub GitHub

Abstract:
Preference-conditioned image generation seeks to adapt generative models to individual users, producing outputs that reflect personal aesthetic choices beyond the given textual prompt. Despite recent progress, existing approaches either fail to capture nuanced user preferences or lack effective mechanisms to encode personalized visual signals. In this work, we propose a multimodal framework that leverages multimodal large language models (MLLMs) to extract rich user representations and inject them into diffusion-based image generation. We train the MLLM with a preference-oriented visual question answering task to capture fine-grained semantic cues. To isolate preference-relevant features, we introduce two complementary probing tasks: inter-user discrimination to distinguish between different users, and intra-user discrimination to separate liked from disliked content. To ensure compatibility with diffusion text encoders, we design a maximum mean discrepancy-based alignment loss that bridges the modality gap while preserving multimodal structure. The resulting embeddings are used to condition the generator, enabling faithful adherence to both prompts and user preferences. Extensive experiments demonstrate that our method substantially outperforms strong baselines in both image quality and preference alignment, highlighting the effectiveness of representation extraction and alignment for personalized generation.

Paperid: 132, https://arxiv.org/pdf/2512.06013.pdf GitHub

Abstract:
In robot learning, Vision Transformers (ViTs) are standard for visual perception, yet most methods discard valuable information by using only the final layer's features. We argue this provides an insufficient representation and propose the Vision Action Transformer (VAT), a novel architecture that is extended from ViT and unlocks the full feature hierarchy of ViT. VAT processes specialized action tokens with visual features across all transformer layers, enabling a deep and progressive fusion of perception and action generation. On a suite of simulated manipulation tasks, VAT achieves a 98.15\% average success rate across four LIBERO benchmarks, establishing a new state-of-the-art by outperforming prior methods like OpenVLA-OFT. Our work presents not only a powerful model for imitation learning but also demonstrates the critical importance of leveraging the complete ''representation trajectory'' of vision models to advance robotic policy. The GitHub URL for the project code is https://github.com/sellerbubble/VAT.

Paperid: 133, https://arxiv.org/pdf/2512.05996.pdf GitHub

Abstract:
Analyzing underwater fish imagery is critical for ecological monitoring but remains difficult due to visual degradation and costly annotations. We introduce FishDetector-R1, a unified MLLM-based framework for fish detection, segmentation, and counting under weak supervision. On the DeepFish dataset, our framework achieves substantial gains over baselines, improving AP by 20% and mIoU by 10%, while reducing MAE by 30% and GAME by 35%. These improvements stem from two key components: a novel detect-to-count prompt that enforces spatially consistent detections and counts, and Reinforcement Learning from Verifiable Reward (RLVR) with a complementary scalable paradigm leveraging sparse point labels. Ablation studies further validate the effectiveness of this reward design. Moreover, the improvement generalizes well to other underwater datasets, confirming strong cross-domain robustness. Overall, FishDetector-R1 provides a reliable and scalable solution for accurate marine visual understanding via weak supervision. The project page for FishDetector-R1 is https://umfieldrobotics.github.io/FishDetector-R1.

Paperid: 134, https://arxiv.org/pdf/2512.05969.pdf GitHub GitHub GitHub GitHub

Abstract:
We show that video generation models could reason now. Testing on tasks such as chess, maze, Sudoku, mental rotation, and Raven's Matrices, leading models such as Sora-2 achieve sixty percent success rates. We establish a robust experimental paradigm centered on the "Task Pair" design. We build a code framework, with 39 models available already, that supports this paradigm and allows for easy scaling - users can add models and tasks efficiently. We show our automated evaluation strongly correlates with human judgment, and therefore this paradigm is highly scalable. We see an opportunity, given the availability of our paradigm, to do reinforcement learning for improving reasoning in video models. You could checkout all of our raw $\href{https://grow-ai-like-a-child.com/video-reason/}{results}$ and our $\href{https://github.com/hokindeng/VMEvalKit}{VMEvalKit}$ codebase.

Paperid: 135, https://arxiv.org/pdf/2512.05965.pdf GitHub

Abstract:
Instruction-based image editing has emerged as a prominent research area, which, benefiting from image generation foundation models, have achieved high aesthetic quality, making instruction-following capability the primary challenge. Existing approaches improve instruction adherence via supervised or reinforcement learning, yet single-turn success rates remain limited due to inherent stochasticity and a lack of deliberation. In this work, we propose a deliberative editing framework to 'think' while they edit, which simulates the human cognitive loop by iteratively executing a Think-while-Edit cycle: Critiquing results and Refining instructions , followed by Repeating the generation until satisfactory. Specifically, we train a single MLLM, EditThinker, to act as the reasoning engine of this framework, which jointly produce the critique score, reasoning process, and refined instructions. We employ reinforcement learning to align the EditThinker's thinking with its editing, thereby generating more targeted instruction improvements. Extensive experiments on four benchmarks demonstrate that our approach significantly improves the instruction-following capability of any image editing model by a large margin. We will release our data construction framework, datasets, and models to benefit the community.

Paperid: 136, https://arxiv.org/pdf/2512.05946.pdf GitHub

Abstract:
Resource allocation remains NP-hard due to combinatorial complexity. While deep reinforcement learning (DRL) methods, such as the Rainbow Deep Q-Network (DQN), improve scalability through prioritized replay and distributional heads, classical function approximators limit their representational power. We introduce Variational Quantum Rainbow DQN (VQR-DQN), which integrates ring-topology variational quantum circuits with Rainbow DQN to leverage quantum superposition and entanglement. We frame the human resource allocation problem (HRAP) as a Markov decision process (MDP) with combinatorial action spaces based on officer capabilities, event schedules, and transition times. On four HRAP benchmarks, VQR-DQN achieves 26.8% normalized makespan reduction versus random baselines and outperforms Double DQN and classical Rainbow DQN by 4.9-13.4%. These gains align with theoretical connections between circuit expressibility, entanglement, and policy quality, demonstrating the potential of quantum-enhanced DRL for large-scale resource allocation. Our implementation is available at: https://github.com/Analytics-Everywhere-Lab/qtrl/.

Paperid: 137, https://arxiv.org/pdf/2512.05941.pdf GitHub

Abstract:
Grounding is a fundamental capability for building graphical user interface (GUI) agents. Although existing approaches rely on large-scale bounding box supervision, they still face various challenges, such as cross-platform generalization, complex layout analysis, and fine-grained element localization. In this paper, we investigate zoom as a strong yet underexplored prior for GUI grounding, and propose a training-free method, ZoomClick. By characterizing four key properties of zoom (i.e., pre-zoom, depth, shrink size, minimal crop size), we unlock its full capabilities for dynamic spatial focusing and adaptive context switching. Experiments demonstrate that our method significantly boosts the performance of both general vision-language and specialized GUI grounding models, achieving state-of-the-art results on several mainstream benchmarks; for example, UI-Venus-72B attains a 73.1% success rate on ScreenSpot-Pro. Furthermore, we present GUIZoom-Bench, a benchmark for evaluating model adaptability to zoom, aiming to inspire future research on improving zoom for further training and test-time scaling in GUI grounding tasks.

Paperid: 138, https://arxiv.org/pdf/2512.05881.pdf GitHub

Abstract:
Traditional physics-informed neural networks (PINNs) do not always satisfy physics based constraints, especially when the constraints include differential operators. Rather, they minimize the constraint violations in a soft way. Strict satisfaction of differential-algebraic equations (DAEs) to embed domain knowledge and first-principles in data-driven models is generally challenging. This is because data-driven models consider the original functions to be black-box whose derivatives can only be obtained after evaluating the functions. We introduce DAE-HardNet, a physics-constrained (rather than simply physics-informed) neural network that learns both the functions and their derivatives simultaneously, while enforcing algebraic as well as differential constraints. This is done by projecting model predictions onto the constraint manifold using a differentiable projection layer. We apply DAE-HardNet to several systems and test problems governed by DAEs, including the dynamic Lotka-Volterra predator-prey system and transient heat conduction. We also show the ability of DAE-HardNet to estimate unknown parameters through a parameter estimation problem. Compared to multilayer perceptrons (MLPs) and PINNs, DAE-HardNet achieves orders of magnitude reduction in the physics loss while maintaining the prediction accuracy. It has the added benefits of learning the derivatives which improves the constrained learning of the backbone neural network prior to the projection layer. For specific problems, this suggests that the projection layer can be bypassed for faster inference. The current implementation and codes are available at https://github.com/SOULS-TAMU/DAE-HardNet.

Paperid: 139, https://arxiv.org/pdf/2512.05832.pdf GitHub

Abstract:
This study examines how interruptions during U.S. Supreme Court oral arguments shape both the semantic content and emotional tone of advocates' speech, with a focus on gendered dynamics in judicial discourse. Using the ConvoKit Supreme Court Corpus (2010-2019), we analyze 12,663 speech chunks from advocate-justice interactions to assess whether interruptions alter the meaning of an advocate's argument and whether interruptions toward female advocates exhibit more negative emotional valence. Semantic shifts are quantified using GloVe-based sentence embeddings, while sentiment is measured through lexicon-based analysis. We find that semantic similarity between pre- and post-interruption speech remains consistently high, suggesting that interruptions do not substantially alter argumentative content. However, interruptions directed at female advocates contain significantly higher levels of negative sentiment. These results deepen empirical understanding of gendered communication in elite institutional settings and demonstrate the value of computational linguistic methods for studying power, discourse, and equity in judicial proceedings.

Paperid: 140, https://arxiv.org/pdf/2512.05830.pdf GitHub

Abstract:
This study focuses on event detection in optical fibers, specifically classifying six events using the Phase-OTDR system. A novel approach is introduced to enhance Phase-OTDR data analysis by transforming 1D data into grayscale images through techniques such as Gramian Angular Difference Field, Gramian Angular Summation Field, and Recurrence Plot. These grayscale images are combined into a multi-channel RGB representation, enabling more robust and adaptable analysis using transfer learning models. The proposed methodology achieves high classification accuracies of 98.84% and 98.24% with the EfficientNetB0 and DenseNet121 models, respectively. A 5-fold cross-validation process confirms the reliability of these models, with test accuracy rates of 99.07% and 98.68%. Using a publicly available Phase-OTDR dataset, the study demonstrates an efficient approach to understanding optical fiber events while reducing dataset size and improving analysis efficiency. The results highlight the transformative potential of image-based analysis in interpreting complex fiber optic sensing data, offering significant advancements in the accuracy and reliability of fiber optic monitoring systems. The codes and the corresponding image-based dataset are made publicly available on GitHub to support further research: https://github.com/miralab-ai/Phase-OTDR-event-detection.

Paperid: 141, https://arxiv.org/pdf/2512.05814.pdf GitHub

Abstract:
Alzheimer's disease (AD) is an irreversible neurodegenerative disorder, and early diagnosis is critical for timely intervention. However, most existing classification frameworks face challenges in multicenter studies, as they often neglect inter-site heterogeneity and lack mechanisms to quantify uncertainty, which limits their robustness and clinical applicability. To address these issues, we proposed Uncertainty-Guided Federated Domain Adaptation (UG-FedDA), a novel multicenter AD classification framework that integrates uncertainty quantification (UQ) with federated domain adaptation to handle cross-site structure magnetic resonance imaging (MRI) heterogeneity under privacy constraints. Our approach extracts multi-template region-of-interest (RoI) features using a self-attention transformer, capturing both regional representations and their interactions. UQ is integrated to guide feature alignment, mitigating source-target distribution shifts by down-weighting uncertain samples. Experiments are conducted on three public datasets: the Alzheimer's Disease Neuroimaging Initiative (ADNI), the Australian Imaging, Biomarkers and Lifestyle study (AIBL), and the Open Access Series of Imaging Studies (OASIS). UG-FedDA achieved consistent cross-domain improvements in accuracy, sensitivity, and area under the ROC curve across three classification tasks: AD vs. normal controls (NC), mild cognitive impairment (MCI) vs. AD, and NC vs. MCI. For NC vs. AD, UG-FedDA achieves accuracies of 90.54%, 89.04%, and 77.78% on ADNI, AIBL and OASIS datasets, respectively. For MCI vs. AD, accuracies are 80.20% (ADNI), 71.91% (AIBL), and 79.73% (OASIS). For NC vs. MCI, results are 76.87% (ADNI), 73.91% (AIBL), and 83.73% (OASIS). These results demonstrate that the proposed framework not only adapts efficiently across multiple sites but also preserves strict privacy.

Paperid: 142, https://arxiv.org/pdf/2512.05809.pdf GitHub

Abstract:
Vision-Language Models (VLMs) remain limited in spatial reasoning tasks that require multi-view understanding and embodied perspective shifts. Recent approaches such as MindJourney attempt to mitigate this gap through test-time scaling where a world model imagines action-conditioned trajectories and a heuristic verifier selects helpful views from such trajectories. In this work, we systematically examine how such test-time verifiers behave across benchmarks, uncovering both their promise and their pitfalls. Our uncertainty-based analyses show that MindJourney's verifier provides little meaningful calibration, and that random scoring often reduces answer entropy equally well, thus exposing systematic action biases and unreliable reward signals. To mitigate these, we introduce a Verification through Spatial Assertions (ViSA) framework that grounds the test-time reward in verifiable, frame-anchored micro-claims. This principled verifier consistently improves spatial reasoning on the SAT-Real benchmark and corrects trajectory-selection biases through more balanced exploratory behavior. However, on the challenging MMSI-Bench, none of the verifiers, including ours, achieve consistent scaling, suggesting that the current world models form an information bottleneck where imagined views fail to enrich fine-grained reasoning. Together, these findings chart the bad, good, and ugly aspects of test-time verification for world-model-based reasoning. Our code is available at https://github.com/chandar-lab/visa-for-mindjourney.

Paperid: 143, https://arxiv.org/pdf/2512.05802.pdf GitHub

Abstract:
Customized text-to-video generation (CTVG) has recently witnessed great progress in generating tailored videos from user-specific text. However, most CTVG methods assume that personalized concepts remain static and do not expand incrementally over time. Additionally, they struggle with forgetting and concept neglect when continuously learning new concepts, including subjects and motions. To resolve the above challenges, we develop a novel Continual Customized Video Diffusion (CCVD) model, which can continuously learn new concepts to generate videos across various text-to-video generation tasks by tackling forgetting and concept neglect. To address catastrophic forgetting, we introduce a concept-specific attribute retention module and a task-aware concept aggregation strategy. They can capture the unique characteristics and identities of old concepts during training, while combining all subject and motion adapters of old concepts based on their relevance during testing. Besides, to tackle concept neglect, we develop a controllable conditional synthesis to enhance regional features and align video contexts with user conditions, by incorporating layer-specific region attention-guided noise estimation. Extensive experimental comparisons demonstrate that our CCVD outperforms existing CTVG models. The code is available at https://github.com/JiahuaDong/CCVD.

Paperid: 144, https://arxiv.org/pdf/2512.05783.pdf GitHub

Abstract:
When depth sensors provide only 5% of needed measurements, reconstructing complete 3D scenes becomes difficult. Autonomous vehicles and robots cannot tolerate the geometric errors that sparse reconstruction introduces. We propose curvature regularization through a discrete Laplacian operator, achieving 18.1% better reconstruction accuracy than standard variational autoencoders. Our contribution challenges an implicit assumption in geometric deep learning: that combining multiple geometric constraints improves performance. A single well-designed regularization term not only matches but exceeds the effectiveness of complex multi-term formulations. The discrete Laplacian offers stable gradients and noise suppression with just 15% training overhead and zero inference cost. Code and models are available at https://github.com/Maryousefi/GeoVAE-3D.

Paperid: 145, https://arxiv.org/pdf/2512.05760.pdf GitHub

Abstract:
Machine intelligence marks the ultimate dream of making machines' intelligence comparable to human beings. While recent progress in Large Language Models (LLMs) show substantial specific skills for a wide array of downstream tasks, they more or less fall shorts in general intelligence. Following correlation between intelligence and system 2 reasoning (slow thinking), in this paper, we aim to answering a worthwhile research question: could machine intelligence such as LLMs be evolved to acquire reasoning ability (not specific skill) just like our human beings? To this end, we propose evolutionary reasoning optimization (ERO) framework which performs survival of the fittest over a population of LLMs to search for individual with strong reasoning ability. Given a reasoning task, ERO first initializes multiple LLMs as a population, after which an evolutionary strategy evolves the population to maximize quantified reasoning score of the best individual. Based on experiments on representative testsuites, we claim two surprising empirical discoveries: i) the latest LLMs such as GPT-5 still show limited system 2 reasoning ability; ii) with simple evolution-loop of ERO, a relatively weak model (Qwen-7B) could be enhanced to emerge powerful reasoning ability. Our project can be accessed at https://github.com/MetaEvo/ERO for reproduction needs.

Paperid: 146, https://arxiv.org/pdf/2512.05693.pdf GitHub

Abstract:
The development of foundation models for embodied intelligence critically depends on access to large-scale, high-quality robot demonstration data. Recent approaches have sought to address this challenge by training on large collections of heterogeneous robotic datasets. However, unlike vision or language data, robotic demonstrations exhibit substantial heterogeneity across embodiments and action spaces as well as other prominent variations such as senor configurations and action control frequencies. The lack of explicit designs for handling such heterogeneity causes existing methods to struggle with integrating diverse factors, thereby limiting their generalization and leading to degraded performance when transferred to new settings. In this paper, we present HiMoE-VLA, a novel vision-language-action (VLA) framework tailored to effectively handle diverse robotic data with heterogeneity. Specifically, we introduce a Hierarchical Mixture-of-Experts (HiMoE) architecture for the action module which adaptively handles multiple sources of heterogeneity across layers and gradually abstracts them into shared knowledge representations. Through extensive experimentation with simulation benchmarks and real-world robotic platforms, HiMoE-VLA demonstrates a consistent performance boost over existing VLA baselines, achieving higher accuracy and robust generalization across diverse robots and action spaces. The code and models are publicly available at https://github.com/ZhiyingDu/HiMoE-VLA.

Paperid: 147, https://arxiv.org/pdf/2512.05683.pdf GitHub

Abstract:
Optical aberrations significantly degrade image quality in microscopy, particularly when imaging deeper into samples. These aberrations arise from distortions in the optical wavefront and can be mathematically represented using Zernike polynomials. Existing methods often address only mild aberrations on limited sample types and modalities, typically treating the problem as a black-box mapping without leveraging the underlying optical physics of wavefront distortions. We propose ZRNet, a physics-informed framework that jointly performs Zernike coefficient prediction and optical image Restoration. We contribute a Zernike Graph module that explicitly models physical relationships between Zernike polynomials based on their azimuthal degrees-ensuring that learned corrections align with fundamental optical principles. To further enforce physical consistency between image restoration and Zernike prediction, we introduce a Frequency-Aware Alignment (FAA) loss, which better aligns Zernike coefficient prediction and image features in the Fourier domain. Extensive experiments on CytoImageNet demonstrates that our approach achieves state-of-the-art performance in both image restoration and Zernike coefficient prediction across diverse microscopy modalities and biological samples with complex, large-amplitude aberrations. Code is available at https://github.com/janetkok/ZRNet.

Paperid: 148, https://arxiv.org/pdf/2512.05672.pdf GitHub

Abstract:
Recent approaches to controllable 4D video generation often rely on fine-tuning pre-trained Video Diffusion Models (VDMs). This dominant paradigm is computationally expensive, requiring large-scale datasets and architectural modifications, and frequently suffers from catastrophic forgetting of the model's original generative priors. Here, we propose InverseCrafter, an efficient inpainting inverse solver that reformulates the 4D generation task as an inpainting problem solved in the latent space. The core of our method is a principled mechanism to encode the pixel space degradation operator into a continuous, multi-channel latent mask, thereby bypassing the costly bottleneck of repeated VAE operations and backpropagation. InverseCrafter not only achieves comparable novel view generation and superior measurement consistency in camera control tasks with near-zero computational overhead, but also excels at general-purpose video inpainting with editing. Code is available at https://github.com/yeobinhong/InverseCrafter.

Paperid: 149, https://arxiv.org/pdf/2512.05669.pdf GitHub

Abstract:
Facial expression recognition is a crucial component in enhancing human-computer interaction and developing emotion-aware systems. Real-time detection and interpretation of facial expressions have become increasingly important for various applications, from user experience personalization to intelligent surveillance systems. This study presents a novel approach to real-time sequential facial expression recognition using deep learning and geometric features. The proposed method utilizes MediaPipe FaceMesh for rapid and accurate facial landmark detection. Geometric features, including Euclidean distances and angles, are extracted from these landmarks. Temporal dynamics are incorporated by analyzing feature differences between consecutive frames, enabling the detection of onset, apex, and offset phases of expressions. For classification, a ConvLSTM1D network followed by multilayer perceptron blocks is employed. The method's performance was evaluated on multiple publicly available datasets, including CK+, Oulu-CASIA (VIS and NIR), and MMI. Accuracies of 93%, 79%, 77%, and 68% were achieved respectively. Experiments with composite datasets were also conducted to assess the model's generalization capabilities. The approach demonstrated real-time applicability, processing approximately 165 frames per second on consumer-grade hardware. This research contributes to the field of facial expression analysis by providing a fast, accurate, and adaptable solution. The findings highlight the potential for further advancements in emotion-aware technologies and personalized user experiences, paving the way for more sophisticated human-computer interaction systems. To facilitate further research in this field, the complete source code for this study has been made publicly available on GitHub: https://github.com/miralab-ai/facial-expression-analysis.

Paperid: 150, https://arxiv.org/pdf/2512.05665.pdf GitHub

Abstract:
Interleaved reasoning paradigms enhance Multimodal Large Language Models (MLLMs) with visual feedback but are hindered by the prohibitive computational cost of repeatedly re-encoding pixel-dense images. A promising alternative, latent visual reasoning, circumvents this bottleneck yet currently forces a critical trade-off: methods either sacrifice precise perceptual modeling by over-compressing features or fail to model dynamic problems due to static, non-interleaved structures. We introduce Interleaved Latent Visual Reasoning (ILVR), a framework that unifies dynamic state evolution with precise perceptual modeling. ILVR interleaves textual generation with latent visual representations that act as specific, evolving cues for subsequent reasoning. To enable this, we employ a self-supervision strategy where a Momentum Teacher Model selectively distills relevant features from helper images into sparse supervision targets. This adaptive selection mechanism guides the model to autonomously generate context-aware visual signals. Extensive experiments on multimodal reasoning benchmarks demonstrate that ILVR significantly outperforms existing approaches, effectively bridging the gap between fine-grained perception and sequential multimodal reasoning.

Paperid: 151, https://arxiv.org/pdf/2512.05635.pdf GitHub

Abstract:
Learned Image Signal Processing (ISP) pipelines offer powerful end-to-end performance but are critically dependent on large-scale paired raw-to-sRGB datasets. This reliance on costly-to-acquire paired data remains a significant bottleneck. To address this challenge, we introduce a novel, unsupervised training framework based on Optimal Transport capable of training arbitrary ISP architectures in both unpaired and paired modes. We are the first to successfully apply Unbalanced Optimal Transport (UOT) for this complex, cross-domain translation task. Our UOT-based framework provides robustness to outliers in the target sRGB data, allowing it to discount atypical samples that would be prohibitively costly to map. A key component of our framework is a novel ``committee of expert discriminators,'' a hybrid adversarial regularizer. This committee guides the optimal transport mapping by providing specialized, targeted gradients to correct specific ISP failure modes, including color fidelity, structural artifacts, and frequency-domain realism. To demonstrate the superiority of our approach, we retrained existing state-of-the-art ISP architectures using our paired and unpaired setups. Our experiments show that while our framework, when trained in paired mode, exceeds the performance of the original paired methods across all metrics, our unpaired mode concurrently achieves quantitative and qualitative performance that rivals, and in some cases surpasses, the original paired-trained counterparts. The code and pre-trained models are available at: https://github.com/gosha20777/EGUOT-ISP.git.

Paperid: 152, https://arxiv.org/pdf/2512.05620.pdf GitHub

Abstract:
Several recently introduced deep learning optimizers utilizing matrix-level preconditioning have shown promising speedups relative to the current dominant optimizer AdamW, particularly in relatively small-scale experiments. However, efforts to validate and replicate their successes have reported mixed results. To better understand the effectiveness of these optimizers at scale, in this work we investigate how to scale preconditioned optimizers via hyperparameter transfer, building on prior works such as $μ$P. We study how the optimal learning rate and weight decay should scale with model width and depth for a wide range of optimizers, including Shampoo, SOAP, and Muon, accounting for the impact of commonly used techniques such as blocking and grafting. We find that scaling the learning rate according to $μ$P improves transfer, but can still suffer from significant finite-width deviations that cause drifting optimal learning rates, which we show can be mitigated by blocking and explicit spectral normalization. For compute-optimal scaling, we find scaling independent weight decay as $1/\mathrm{width}$ is nearly optimal across optimizers. Applying these scaling rules, we show Muon and Shampoo consistently achieve $1.4\times$ and $1.3\times$ speedup over AdamW for training Llama-architecture language models of sizes ranging from $190$M to $1.4$B, whereas the speedup vanishes rapidly with scale under incorrect scaling. Based on these results and further ablations, we argue that studying optimal hyperparameter transfer is essential for reliably comparing optimizers at scale given a realistic tuning budget.

Paperid: 153, https://arxiv.org/pdf/2512.05619.pdf GitHub

Abstract:
Partial Maximum Satisfiability (PMS) and Weighted Partial Maximum Satisfiability (WPMS) generalize Maximum Satisfiability (MaxSAT), with broad real-world applications. Recent advances in Stochastic Local Search (SLS) algorithms for solving (W)PMS have mainly focused on designing clause weighting schemes. However, existing methods often fail to adequately distinguish between PMS and WPMS, typically employing uniform update strategies for clause weights and overlooking critical structural differences between the two problem types. In this work, we present a novel clause weighting scheme that, for the first time, updates the clause weights of PMS and WPMS instances according to distinct conditions. This scheme also introduces a new initialization method, which better accommodates the unique characteristics of both instance types. Furthermore, we propose a decimation method that prioritizes satisfying unit and hard clauses, effectively complementing our proposed clause weighting scheme. Building on these methods, we develop a new SLS solver for (W)PMS named DeepDist. Experimental results on benchmarks from the anytime tracks of recent MaxSAT Evaluations show that DeepDist outperforms state-of-the-art SLS solvers. Notably, a hybrid solver combining DeepDist with TT-Open-WBO-Inc surpasses the performance of the MaxSAT Evaluation 2024 winners, SPB-MaxSAT-c-Band and SPB-MaxSAT-c-FPS, highlighting the effectiveness of our approach. The code is available at https://github.com/jmhmaxsat/DeepDist

Paperid: 154, https://arxiv.org/pdf/2512.05613.pdf GitHub

Abstract:
Cross-Domain Few-Shot Semantic Segmentation (CD-FSS) seeks to segment unknown classes in unseen domains using only a few annotated examples. This setting is inherently challenging: source and target domains exhibit substantial distribution shifts, label spaces are disjoint, and support images are scarce--making standard episodic methods unreliable and computationally demanding at test time. To address these constraints, we propose DistillFSS, a framework that embeds support-set knowledge directly into a model's parameters through a teacher--student distillation process. By internalizing few-shot reasoning into a dedicated layer within the student network, DistillFSS eliminates the need for support images at test time, enabling fast, lightweight inference, while allowing efficient extension to novel classes in unseen domains through rapid teacher-driven specialization. Combined with fine-tuning, the approach scales efficiently to large support sets and significantly reduces computational overhead. To evaluate the framework under realistic conditions, we introduce a new CD-FSS benchmark spanning medical imaging, industrial inspection, and remote sensing, with disjoint label spaces and variable support sizes. Experiments show that DistillFSS matches or surpasses state-of-the-art baselines, particularly in multi-class and multi-shot scenarios, while offering substantial efficiency gains. The code is available at https://github.com/pasqualedem/DistillFSS.

Paperid: 155, https://arxiv.org/pdf/2512.05578.pdf GitHub

Abstract:
Hyperspectral imaging is an advanced technique for precisely identifying and analyzing materials or objects. However, its integration with robotic grasping systems has so far been explored due to the deployment complexities and prohibitive costs. Within this paper, we introduce a novel hyperspectral imaging-guided robotic grasping system. The system consists of PRISM (Polyhedral Reflective Imaging Scanning Mechanism) and the SpectralGrasp framework. PRISM is designed to enable high-precision, distortion-free hyperspectral imaging while simplifying system integration and costs. SpectralGrasp generates robotic grasping strategies by effectively leveraging both the spatial and spectral information from hyperspectral images. The proposed system demonstrates substantial improvements in both textile recognition compared to human performance and sorting success rate compared to RGB-based methods. Additionally, a series of comparative experiments further validates the effectiveness of our system. The study highlights the potential benefits of integrating hyperspectral imaging with robotic grasping systems, showcasing enhanced recognition and grasping capabilities in complex and dynamic environments. The project is available at: https://zainzh.github.io/PRISM.

Paperid: 156, https://arxiv.org/pdf/2512.05576.pdf GitHub GitHub

Abstract:
Current clinical agent built on small LLMs, such as TxAgent suffer from a \textit{Context Utilization Failure}, where models successfully retrieve biomedical evidence due to supervised finetuning but fail to ground their diagnosis in that information. In this work, we propose the Executor-Analyst Framework, a modular architecture that decouples the syntactic precision of tool execution from the semantic robustness of clinical reasoning. By orchestrating specialized TxAgents (Executors) with long-context foundation models (Analysts), we mitigate the reasoning deficits observed in monolithic models. Beyond simple modularity, we demonstrate that a Stratified Ensemble strategy significantly outperforms global pooling by preserving evidentiary diversity, effectively addressing the information bottleneck. Furthermore, our stress tests reveal critical scaling insights: (1) a \textit{Context-Performance Paradox}, where extending reasoning contexts beyond 12k tokens introduces noise that degrades accuracy; and (2) the \textit{Curse of Dimensionality} in action spaces, where expanding toolsets necessitates hierarchical retrieval strategies. Crucially, our approach underscores the potential of training-free architectural engineering, achieving state-of-the-art performance on CURE-Bench without the need for expensive end-to-end finetuning. This provides a scalable, agile foundation for the next generation of trustworthy AI-driven therapeutics. Code has been released on https://github.com/June01/CureAgent.

Paperid: 157, https://arxiv.org/pdf/2512.05530.pdf GitHub

Abstract:
Recently, multimodal large language models (MLLMs) have been widely applied to reasoning tasks. However, they suffer from limited multi-rationale semantic modeling, insufficient logical robustness, and are susceptible to misleading interpretations in complex scenarios. Therefore, we propose a Multi-rationale INtegrated Discriminative (MIND) reasoning framework, which is designed to endow MLLMs with human-like cognitive abilities of "Understand -> Rethink -> Correct", and achieves a paradigm evolution from passive imitation-based reasoning to active discriminative reasoning. Specifically, we introduce a Rationale Augmentation and Discrimination (RAD) paradigm, which automatically and efficiently expands existing datasets by generating diverse rationales, providing a unified and extensible data foundation. Meanwhile, we design a Progressive Two-stage Correction Learning (P2CL) strategy. The first phase enhances multi-rationale positive learning, while the second phase enables active logic discrimination and correction. In addition, to mitigate representation entanglement in the multi-rationale semantic space, we propose a Multi-rationale Contrastive Alignment (MCA) optimization strategy, which achieves semantic aggregation of correct reasoning and boundary separation of incorrect reasoning. Extensive experiments demonstrate that the proposed MIND reasoning framework achieves state-of-the-art (SOTA) performance on multiple public datasets covering scientific, commonsense, and mathematical scenarios. It provides a new perspective for advancing MLLMs towards higher levels of cognitive intelligence. Our code is available at https://github.com/YuChuang1205/MIND

Paperid: 158, https://arxiv.org/pdf/2512.05524.pdf GitHub

Abstract:
Spatio-temporal scene graph generation (ST-SGG) aims to model objects and their evolving relationships across video frames, enabling interpretable representations for downstream reasoning tasks such as video captioning and visual question answering. Despite recent advancements in DETR-style single-stage ST-SGG models, they still suffer from several key limitations. First, while these models rely on attention-based learnable queries as a core component, these learnable queries are semantically uninformed and instance-agnostically initialized. Second, these models rely exclusively on unimodal visual features for predicate classification. To address these challenges, we propose VOST-SGG, a VLM-aided one-stage ST-SGG framework that integrates the common sense reasoning capabilities of vision-language models (VLMs) into the ST-SGG pipeline. First, we introduce the dual-source query initialization strategy that disentangles what to attend to from where to attend, enabling semantically grounded what-where reasoning. Furthermore, we propose a multi-modal feature bank that fuses visual, textual, and spatial cues derived from VLMs for improved predicate classification. Extensive experiments on the Action Genome dataset demonstrate that our approach achieves state-of-the-art performance, validating the effectiveness of integrating VLM-aided semantic priors and multi-modal features for ST-SGG. We will release the code at https://github.com/LUNAProject22/VOST.

Paperid: 159, https://arxiv.org/pdf/2512.05515.pdf GitHub

Abstract:
Multimodal sentiment analysis (MSA) integrates various modalities, such as text, image, and audio, to provide a more comprehensive understanding of sentiment. However, effective MSA is challenged by alignment and fusion issues. Alignment requires synchronizing both temporal and semantic information across modalities, while fusion involves integrating these aligned features into a unified representation. Existing methods often address alignment or fusion in isolation, leading to limitations in performance and efficiency. To tackle these issues, we propose a novel framework called Dual-stream Alignment with Hierarchical Bottleneck Fusion (DashFusion). Firstly, dual-stream alignment module synchronizes multimodal features through temporal and semantic alignment. Temporal alignment employs cross-modal attention to establish frame-level correspondences among multimodal sequences. Semantic alignment ensures consistency across the feature space through contrastive learning. Secondly, supervised contrastive learning leverages label information to refine the modality features. Finally, hierarchical bottleneck fusion progressively integrates multimodal information through compressed bottleneck tokens, which achieves a balance between performance and computational efficiency. We evaluate DashFusion on three datasets: CMU-MOSI, CMU-MOSEI, and CH-SIMS. Experimental results demonstrate that DashFusion achieves state-of-the-art performance across various metrics, and ablation studies confirm the effectiveness of our alignment and fusion techniques. The codes for our experiments are available at https://github.com/ultramarineX/DashFusion.

Paperid: 160, https://arxiv.org/pdf/2512.05513.pdf GitHub

Abstract:
Large Video-Language Models (Video-LMs) have achieved impressive progress in multimodal understanding, yet their reasoning remains weakly grounded in space and time. We present Know-Show, a new benchmark designed to evaluate spatio-temporal grounded reasoning, the ability of a model to reason about actions and their semantics while simultaneously grounding its inferences in visual and temporal evidence. Know-Show unifies reasoning and localization within a single evaluation framework consisting of five complementary scenarios across spatial (person, object, person-object, and hand-object) and temporal dimensions. Built from Charades, Action Genome, and Ego4D with 2.5K human-authored questions, the benchmark exposes significant gaps between current Video-LMs and human reasoning. To bridge this gap, we propose GRAM, a training-free plug-in that augments Video-LMs with fine-grained grounding through attention-based video token selection and explicit timestamp encoding. Extensive experiments across open and closed Video-LMs (Qwen, VideoLLaVA, GPT-4o, and Gemini, etc.) reveal that existing models struggle to "show what they know" and vice versa, especially in fine-grained hand-object interactions. Know-Show establishes a unified standard for assessing grounded reasoning in video-language understanding and provides insights toward developing interpretable and reliable multimodal reasoning systems. We will release the dataset and the code at https://github.com/LUNAProject22/Know-Show.

Paperid: 161, https://arxiv.org/pdf/2512.05511.pdf GitHub

Abstract:
While large-scale visual foundation models (VFMs) exhibit strong generalization across diverse visual domains, their potential for single-frame infrared small target (SIRST) detection remains largely unexplored. To fill this gap, we systematically introduce the frozen representations from VFMs into the SIRST task for the first time and propose a Foundation-Driven Efficient Paradigm (FDEP), which can seamlessly adapt to existing encoder-decoder-based methods and significantly improve accuracy without additional inference overhead. Specifically, a Semantic Alignment Modulation Fusion (SAMF) module is designed to achieve dynamic alignment and deep fusion of the global semantic priors from VFMs with task-specific features. Meanwhile, to avoid the inference time burden introduced by VFMs, we propose a Collaborative Optimization-based Implicit Self-Distillation (CO-ISD) strategy, which enables implicit semantic transfer between the main and lightweight branches through parameter sharing and synchronized backpropagation. In addition, to unify the fragmented evaluation system, we construct a Holistic SIRST Evaluation (HSE) metric that performs multi-threshold integral evaluation at both pixel-level confidence and target-level robustness, providing a stable and comprehensive basis for fair model comparison. Extensive experiments demonstrate that the SIRST detection networks equipped with our FDEP framework achieve state-of-the-art (SOTA) performance on multiple public datasets. Our code is available at https://github.com/YuChuang1205/FDEP-Framework

Paperid: 162, https://arxiv.org/pdf/2512.05505.pdf GitHub

Abstract:
We present a method to compute explicit solutions of parametric Generalized Nash Equilibrium (GNE) problems with convex quadratic cost functions and linear coupling and local constraints. Assuming the parameters only enter the linear terms of the cost functions and constraint right-hand sides, we provide the exact multiparametric solution of the GNE problem. Such a solution enables (i) minimal real-time computation, (ii) inherent interpretability, explainability, and exact enumeration of all multiple equilibria, (iii) determine desired GNE solution types in the case of infinitely-many equilibria, and (iv) zero-shot updates of the GNE solution due to changes of constraint right-hand sides and/or linear costs. In line with explicit Model Predictive Control (MPC) approaches, we apply our method to solve game-theoretic MPC (Receding Horizon Games) explicitly, comparing performance against centralized solvers in a battery charging game and in a toy two-mass spring system control problem. A Python implementation of the algorithms presented in this paper is available on https://github.com/bemporad/nash_mpqp.

Paperid: 163, https://arxiv.org/pdf/2512.05482.pdf GitHub

Abstract:
Rare-object detection remains a challenging task in autonomous driving systems, particularly when relying solely on point cloud data. Although Vision-Language Models (VLMs) exhibit strong capabilities in image understanding, their potential to enhance 3D object detection through intelligent data mining has not been fully explored. This paper proposes a novel cross-modal framework that leverages 2D VLMs to identify and mine rare objects from driving scenes, thereby improving 3D object detection performance. Our approach synthesizes complementary techniques such as object detection, semantic feature extraction, dimensionality reduction, and multi-faceted outlier detection into a cohesive, explainable pipeline that systematically identifies rare but critical objects in driving scenes. By combining Isolation Forest and t-SNE-based outlier detection methods with concept-based filtering, the framework effectively identifies semantically meaningful rare objects. A key strength of this approach lies in its ability to extract and annotate targeted rare object concepts such as construction vehicles, motorcycles, and barriers. This substantially reduces the annotation burden and focuses only on the most valuable training samples. Experiments on the nuScenes dataset demonstrate that this concept-guided data mining strategy enhances the performance of 3D object detection models while utilizing only a fraction of the training data, with particularly notable improvements for challenging object categories such as trailers and bicycles compared with the same amount of random data. This finding has substantial implications for the efficient curation of datasets in safety-critical autonomous systems.

Paperid: 164, https://arxiv.org/pdf/2512.05481.pdf GitHub

Abstract:
Recently, Multi-Contrast MR Reconstruction (MCMR) has emerged as a hot research topic that leverages high-quality auxiliary modalities to reconstruct undersampled target modalities of interest. However, existing methods often struggle to generalize across different k-space undersampling patterns, requiring the training of a separate model for each specific pattern, which limits their practical applicability. To address this challenge, we propose UniFS, a Unified Frequency-Spatial Fusion model designed to handle multiple k-space undersampling patterns for MCMR tasks without any need for retraining. UniFS integrates three key modules: a Cross-Modal Frequency Fusion module, an Adaptive Mask-Based Prompt Learning module, and a Dual-Branch Complementary Refinement module. These modules work together to extract domain-invariant features from diverse k-space undersampling patterns while dynamically adapt to their own variations. Another limitation of existing MCMR methods is their tendency to focus solely on spatial information while neglect frequency characteristics, or extract only shallow frequency features, thus failing to fully leverage complementary cross-modal frequency information. To relieve this issue, UniFS introduces an adaptive prompt-guided frequency fusion module for k-space learning, significantly enhancing the model's generalization performance. We evaluate our model on the BraTS and HCP datasets with various k-space undersampling patterns and acceleration factors, including previously unseen patterns, to comprehensively assess UniFS's generalizability. Experimental results across multiple scenarios demonstrate that UniFS achieves state-of-the-art performance. Our code is available at https://github.com/LIKP0/UniFS.

Paperid: 165, https://arxiv.org/pdf/2512.05469.pdf GitHub

Abstract:
Ensemble models often achieve higher accuracy than single learners, but their ability to maintain small generalization gaps is not always well understood. This study examines how ensembles balance accuracy and overfitting across four tabular classification tasks: Breast Cancer, Heart Disease, Pima Diabetes, and Credit Card Fraud. Using repeated stratified cross validation with statistical significance testing, we compare linear models, a single decision tree, and nine ensemble methods. The results show that ensembles can reach high accuracy without large gaps by reducing variance through averaging or controlled boosting. On nearly linear and clean data, linear models already generalize well and ensembles offer little additional benefit. On datasets with meaningful nonlinear structure, tree based ensembles increase test accuracy by 5 to 7 points while keeping gaps below 3 percent. On noisy or highly imbalanced datasets, ensembles remain competitive but require regularization to avoid fitting noise or majority class patterns. We also compute simple dataset complexity indicators, such as linearity score, Fisher ratio, and noise estimate, which explain when ensembles are likely to control variance effectively. Overall, the study provides a clear view of how and when ensembles maintain high accuracy while keeping overfitting low, offering practical guidance for model selection in real world tabular applications.

Paperid: 166, https://arxiv.org/pdf/2512.05468.pdf GitHub

Abstract:
Many industrial sectors have been using of machine learning at inference mode on edge devices. Future directions show that training on edge devices is promising due to improvements in semiconductor performance. Wireless Ad Hoc Federated Learning (WAFL) has been proposed as a promising approach for collaborative learning with device-to-device communication among edges. In particular, WAFL with Vision Transformer (WAFL-ViT) has been tested on image recognition tasks with the UTokyo Building Recognition Dataset (UTBR). Since WAFL-ViT is a mission-oriented sensor system, it is essential to construct specific datasets by each mission. In our work, we have developed the Chulalongkorn University Building Recognition Dataset (CUBR), which is specialized for Chulalongkorn University as a case study in Thailand. Additionally, our results also demonstrate that training on WAFL scenarios achieves better accuracy than self-training scenarios. Dataset is available in https://github.com/jo2lxq/wafl/.

Paperid: 167, https://arxiv.org/pdf/2512.05459.pdf GitHub

Abstract:
Large language models (LLMs) have presented outstanding performance in code generation and completion. However, fine-tuning these models on private datasets can raise privacy and proprietary concerns, such as the leakage of sensitive personal information. Differentially private (DP) code generation provides theoretical guarantees for protecting sensitive code by generating synthetic datasets that preserve statistical properties while reducing privacy leakage concerns. However, DP code generation faces significant challenges due to the strict syntactic dependencies and the privacy-utility trade-off. We propose PrivCode, the first DP synthesizer specifically designed for code datasets. It incorporates a two-stage framework to improve both privacy and utility. In the first stage, termed "privacy-sanitizing", PrivCode generates DP-compliant synthetic code by training models using DP-SGD while introducing syntactic information to preserve code structure. The second stage, termed "utility-boosting", fine-tunes a larger pre-trained LLM on the synthetic privacy-free code to mitigate the utility loss caused by DP, enhancing the utility of the generated code. Extensive experiments on four LLMs show that PrivCode generates higher-utility code across various testing tasks under four benchmarks. The experiments also confirm its ability to protect sensitive data under varying privacy budgets. We provide the replication package at the anonymous link.

Paperid: 168, https://arxiv.org/pdf/2512.05453.pdf GitHub

Abstract:
Motivated by the challenges of implementing policy-based data access control (PBAC) under multiple simultaneously applicable compliance frameworks, we present Parajudica, an open, modular, and extensible RDF/SPARQL-based rule system for evaluating context-dependent data compliance status. We demonstrate the utility of this resource and accompanying metamodel through application to existing legal frameworks and industry standards, offering insights for comparative framework analysis. Applications include compliance policy enforcement, compliance monitoring, data discovery, and risk assessment.

Paperid: 169, https://arxiv.org/pdf/2512.05422.pdf GitHub

Abstract:
Unified multimodal models significantly improve visual generation by combining vision-language models (VLMs) with diffusion models. However, existing methods struggle to fully balance sufficient interaction and flexible implementation due to vast representation difference. Considering abundant and hierarchical information in VLM's layers from low-level details to high-level semantics, we propose \textbf{ParaUni}. It extracts features from variants VLM's layers in a \textbf{Para}llel way for comprehensive information interaction and retains a flexible separation architecture to enhance generation in \textbf{Uni}fied multimodal model. Concretely, visual features from all VLM's layers are fed in parallel into a Layer Integration Module (LIM), which efficiently integrates fine-grained details and semantic abstractions and provides the fused representation as a condition to the diffusion model. To further enhance performance, we reveal that these hierarchical layers respond unequally to different rewards in Reinforcement Learning (RL). Crucially, we design a Layer-wise Dynamic Adjustment Mechanism (LDAM) to facilitate multiple reward improvements that aligns the hierarchical properties of these layers using RL. Extensive experiments show ParaUni leverages complementary multi-layer features to substantially improve generation quality and shows strong potential for multiple reward advances during RL stages. Code is available at https://github.com/JosephTiTan/ParaUni.

Paperid: 170, https://arxiv.org/pdf/2512.05418.pdf GitHub

Abstract:
UAV-based autonomous forestry operations require rapid and precise tree branch segmentation for safe navigation and automated pruning across varying pixel resolutions and operational conditions. We evaluate different deep learning methods at three resolutions (256x256, 512x512, 1024x1024) using the Urban Street Tree Dataset, employing standard metrics (IoU, Dice) and specialized measures including Thin Structure IoU (TS-IoU) and Connectivity Preservation Rate (CPR). Among 22 configurations tested, U-Net with MiT-B4 backbone achieves strong performance at 256x256. At 512x512, MiT-B4 leads in IoU, Dice, TS-IoU, and Boundary-F1. At 1024x1024, U-Net+MiT-B3 shows the best validation performance for IoU/Dice and precision, while U-Net++ excels in boundary quality. PSPNet provides the most efficient option (2.36/9.43/37.74 GFLOPs) with 25.7/19.6/11.8 percentage point IoU reductions compared to top performers at respective resolutions. These results establish multi-resolution benchmarks for accuracy-efficiency trade-offs in embedded forestry systems. Implementation is available at https://github.com/BennyLinntu/PerformanceTreeBranchSegmentation.

Paperid: 171, https://arxiv.org/pdf/2512.05402.pdf GitHub

Abstract:
Bitcoin mining hardware acquisition requires strategic timing due to volatile markets, rapid technological obsolescence, and protocol-driven revenue cycles. Despite mining's evolution into a capital-intensive industry, there is little guidance on when to purchase new Application-Specific Integrated Circuit (ASIC) hardware, and no prior computational frameworks address this decision problem. We address this gap by formulating hardware acquisition as a time series classification task, predicting whether purchasing ASIC machines yields profitable (Return on Investment (ROI) >= 1), marginal (0 < ROI < 1), or unprofitable (ROI <= 0) returns within one year. We propose MineROI-Net, an open source Transformer-based architecture designed to capture multi-scale temporal patterns in mining profitability. Evaluated on data from 20 ASIC miners released between 2015 and 2024 across diverse market regimes, MineROI-Net outperforms LSTM-based and TSLANet baselines, achieving 83.7% accuracy and 83.1% macro F1-score. The model demonstrates strong economic relevance, achieving 93.6% precision in detecting unprofitable periods and 98.5% precision for profitable ones, while avoiding misclassification of profitable scenarios as unprofitable and vice versa. These results indicate that MineROI-Net offers a practical, data-driven tool for timing mining hardware acquisitions, potentially reducing financial risk in capital-intensive mining operations. The model is available through: https://github.com/AMAAI-Lab/MineROI-Net.

Paperid: 172, https://arxiv.org/pdf/2512.05398.pdf GitHub

Abstract:
Estimating accurate camera poses, 3D scene geometry, and object motion from in-the-wild videos is a long-standing challenge for classical structure from motion pipelines due to the presence of dynamic objects. Recent learning-based methods attempt to overcome this challenge by training motion estimators to filter dynamic objects and focus on the static background. However, their performance is largely limited by the availability of large-scale motion segmentation datasets, resulting in inaccurate segmentation and, therefore, inferior structural 3D understanding. In this work, we introduce the Dynamic Prior (\ourmodel) to robustly identify dynamic objects without task-specific training, leveraging the powerful reasoning capabilities of Vision-Language Models (VLMs) and the fine-grained spatial segmentation capacity of SAM2. \ourmodel can be seamlessly integrated into state-of-the-art pipelines for camera pose optimization, depth reconstruction, and 4D trajectory estimation. Extensive experiments on both synthetic and real-world videos demonstrate that \ourmodel not only achieves state-of-the-art performance on motion segmentation, but also significantly improves accuracy and robustness for structural 3D understanding.

Paperid: 173, https://arxiv.org/pdf/2512.05394.pdf GitHub

Abstract:
Latent diffusion models pair VAEs with diffusion backbones, and the structure of VAE latents strongly influences the difficulty of diffusion training. However, existing video VAEs typically focus on reconstruction fidelity, overlooking latent structure. We present a statistical analysis of video VAE latent spaces and identify two spectral properties essential for diffusion training: a spatio-temporal frequency spectrum biased toward low frequencies, and a channel-wise eigenspectrum dominated by a few modes. To induce these properties, we propose two lightweight, backbone-agnostic regularizers: Local Correlation Regularization and Latent Masked Reconstruction. Experiments show that our Spectral-Structured VAE (SSVAE) achieves a $3\times$ speedup in text-to-video generation convergence and a 10\% gain in video reward, outperforming strong open-source VAEs. The code is available at https://github.com/zai-org/SSVAE.

Paperid: 174, https://arxiv.org/pdf/2512.05392.pdf GitHub

Abstract:
Sequential quadratic programming (SQP) methods have been remarkably successful in solving a broad range of nonlinear optimization problems. These methods iteratively construct and solve quadratic programming (QP) subproblems to compute directions that converge to a local minimum. While numerous open-source and commercial SQP algorithms are available, their implementations lack the transparency and modularity necessary to adapt and fine-tune them for specific applications or to swap out different modules to create a new optimizer. To address this gap, we present OpenSQP, a modular and reconfigurable SQP algorithm implemented in Python that achieves robust performance comparable to leading algorithms. We implement OpenSQP in a manner that allows users to easily modify or replace components such as merit functions, line search procedures, Hessian approximations, and QP solvers. This flexibility enables the creation of tailored variants of the algorithm for specific needs. To demonstrate reliability, we present numerical results using the standard configuration of OpenSQP that employs a smooth augmented Lagrangian merit function for the line search and a quasi-Newton BFGS method for approximating the Hessians. We benchmark this configuration on a comprehensive set of problems from the CUTEst test suite. The results demonstrate performance that is competitive with proven nonlinear optimization algorithms such as SLSQP, SNOPT, and IPOPT.

Paperid: 175, https://arxiv.org/pdf/2512.05362.pdf GitHub

Abstract:
Lifting Structure-from-Motion (SfM) information from sequential and non-sequential image data is a time-consuming and computationally expensive task. In addition to this, the majority of publicly available data is unfit for processing due to inadequate camera pose variation, obscuring scene elements, and noisy data. To solve this problem, we introduce PoolNet, a versatile deep learning framework for frame-level and scene-level validation of in-the-wild data. We demonstrate that our model successfully differentiates SfM ready scenes from those unfit for processing while significantly undercutting the amount of time state of the art algorithms take to obtain structure-from-motion data.

Paperid: 176, https://arxiv.org/pdf/2512.05354.pdf GitHub

Abstract:
The rise of 3D Gaussian Splatting has revolutionized photorealistic 3D asset creation, yet a critical gap remains for their interactive refinement and editing. Existing approaches based on diffusion or optimization are ill-suited for this task, as they are often prohibitively slow, destructive to the original asset's identity, or lack the precision for fine-grained control. To address this, we introduce \ourmethod, a state-aware feedforward model that enables continuous editing of 3D Gaussian assets from user-provided 2D view(s). Our method directly predicts updates to the attributes of a compact, feature-rich Gaussian representation and leverages Test-Time Training to create a state-aware, iterative workflow. The versatility of our approach allows a single architecture to perform diverse tasks, including high-fidelity local detail refinement, local paint-over, and consistent global recoloring, all at interactive speeds, paving the way for fluid and intuitive 3D content authoring.

Paperid: 177, https://arxiv.org/pdf/2512.05277.pdf GitHub GitHub GitHub

Abstract:
Temporal understanding in autonomous driving (AD) remains a significant challenge, even for recent state-of-the-art (SoTA) Vision-Language Models (VLMs). Prior work has introduced datasets and benchmarks aimed at improving temporal reasoning, but these have emphasized other video content, including sports, cooking, and movies. No existing benchmark focuses exclusively on the unique challenges of temporal understanding in ego-centric AD footage. To fill this gap, the Temporal Understanding in Autonomous Driving (TAD) benchmark is presented, which evaluates VLMs' ability to capture the dynamic relationships between actions in AD. TAD comprises nearly 6,000 question-answer (QA) pairs, spanning 7 human-designed tasks. In addition, an evaluation is performed that consists of 9 closed- and open-source generalist models as well as SoTA AD specialist models. When applied to TAD, current SoTA models demonstrated substandard accuracies, largely due to imperfect fine-grained motion understanding. To improve motion understanding and overall accuracy on TAD, two novel training-free solutions are proposed: Scene-CoT, that leverages Chain-of-Thought (CoT) and TCogMap, which incorporates an ego-centric temporal cognitive map. The proposed approaches are integrated with existing VLMs and improve average accuracy on TAD by up to 17.72%. By introducing TAD, benchmarking multiple SoTA models, and proposing effective enhancements, this work aims to catalyze future research on temporal understanding in AD. The benchmark and evaluation code are available at \href{https://huggingface.co/datasets/vbdai/TAD}{Hugging Face} and \href{https://github.com/vbdi/tad_bench}{Github}, respectively.

Paperid: 178, https://arxiv.org/pdf/2512.05272.pdf GitHub

Abstract:
Scenes in the real world are often composed of several static and dynamic objects. Capturing their 4-dimensional structures, composition and spatio-temporal configuration in-the-wild, though extremely interesting, is equally hard. Therefore, existing works often focus on one object at a time, while relying on some category-specific parametric shape model for dynamic objects. This can lead to inconsistent scene configurations, in addition to being limited to the modeled object categories. We propose COM4D (Compositional 4D), a method that consistently and jointly predicts the structure and spatio-temporal configuration of 4D/3D objects using only static multi-object or dynamic single object supervision. We achieve this by a carefully designed training of spatial and temporal attentions on 2D video input. The training is disentangled into learning from object compositions on the one hand, and single object dynamics throughout the video on the other, thus completely avoiding reliance on 4D compositional training data. At inference time, our proposed attention mixing mechanism combines these independently learned attentions, without requiring any 4D composition examples. By alternating between spatial and temporal reasoning, COM4D reconstructs complete and persistent 4D scenes with multiple interacting objects directly from monocular videos. Furthermore, COM4D provides state-of-the-art results in existing separate problems of 4D object and composed 3D reconstruction despite being purely data-driven.

Paperid: 179, https://arxiv.org/pdf/2512.05262.pdf GitHub

Abstract:
Dafny is a verification-aware programming language that comes with a compiler and static program verifier. However, neither the compiler nor the verifier is proved correct; in fact, soundness bugs have been found in both tools. This paper shows that the aforementioned Dafny tools can be developed with foundational correctness guarantees. We present a functional big-step semantics for an imperative subset of Dafny and, based on this semantics, a verified verification condition generator (VCG) and a verified compiler for Dafny. The subset of Dafny we have formalized includes mutually recursive method calls, while loops, and arrays -- these language features are significant enough to cover challenging examples such as McCarthy's 91 function and array-based programs that are used when teaching Dafny. The verified VCG allows one to prove functional correctness of annotated Dafny programs, while the verified compiler can be used to compile verified Dafny programs to CakeML programs. From there, one can obtain executable machine code via the (already verified) CakeML compiler, all while provably maintaining the functional correctness guarantees that were proved for the source-level Dafny programs. Our work has been mechanized in the HOL4 theorem prover.

Paperid: 180, https://arxiv.org/pdf/2512.05245.pdf GitHub

Abstract:
Accurate prediction of protein function is essential for elucidating molecular mechanisms and advancing biological and therapeutic discovery. Yet experimental annotation lags far behind the rapid growth of protein sequence data. Computational approaches address this gap by associating proteins with Gene Ontology (GO) terms, which encode functional knowledge through hierarchical relations and textual definitions. However, existing models often emphasize one modality over the other, limiting their ability to generalize, particularly to unseen or newly introduced GO terms that frequently arise as the ontology evolves, and making the previously trained models outdated. We present STAR-GO, a Transformer-based framework that jointly models the semantic and structural characteristics of GO terms to enhance zero-shot protein function prediction. STAR-GO integrates textual definitions with ontology graph structure to learn unified GO representations, which are processed in hierarchical order to propagate information from general to specific terms. These representations are then aligned with protein sequence embeddings to capture sequence-function relationships. STAR-GO achieves state-of-the-art performance and superior zero-shot generalization, demonstrating the utility of integrating semantics and structure for robust and adaptable protein function prediction. Code is available at https://github.com/boun-tabi-lifelu/stargo.

Paperid: 181, https://arxiv.org/pdf/2512.05137.pdf GitHub

Abstract:
Vision-Language Models (VLMs) have advanced multimodal understanding, yet still struggle when targets are embedded in cluttered backgrounds requiring figure-ground segregation. To address this, we introduce ChromouVQA, a large-scale, multi-task benchmark based on Ishihara-style chromatic camouflaged images. We extend classic dot plates with multiple fill geometries and vary chromatic separation, density, size, occlusion, and rotation, recording full metadata for reproducibility. The benchmark covers nine vision-question-answering tasks, including recognition, counting, comparison, and spatial reasoning. Evaluations of humans and VLMs reveal large gaps, especially under subtle chromatic contrast or disruptive geometric fills. We also propose a model-agnostic contrastive recipe aligning silhouettes with their camouflaged renderings, improving recovery of global shapes. ChromouVQA provides a compact, controlled benchmark for reproducible evaluation and extension. Code and dataset are available at https://github.com/Chromou-VQA-Benchmark/Chromou-VQA.

Paperid: 182, https://arxiv.org/pdf/2512.05134.pdf GitHub

Abstract:
Diffusion models deliver high-fidelity synthesis but remain slow due to iterative sampling. We empirically observe there exists feature invariance in deterministic sampling, and present InvarDiff, a training-free acceleration method that exploits the relative temporal invariance across timestep-scale and layer-scale. From a few deterministic runs, we compute a per-timestep, per-layer, per-module binary cache plan matrix and use a re-sampling correction to avoid drift when consecutive caches occur. Using quantile-based change metrics, this matrix specifies which module at which step is reused rather than recomputed. The same invariance criterion is applied at the step scale to enable cross-timestep caching, deciding whether an entire step can reuse cached results. During inference, InvarDiff performs step-first and layer-wise caching guided by this matrix. When applied to DiT and FLUX, our approach reduces redundant compute while preserving fidelity. Experiments show that InvarDiff achieves $2$-$3\times$ end-to-end speed-ups with minimal impact on standard quality metrics. Qualitatively, we observe almost no degradation in visual quality compared with full computations.

Paperid: 183, https://arxiv.org/pdf/2512.05131.pdf GitHub

Abstract:
Active 3D reconstruction enables an agent to autonomously select viewpoints to efficiently obtain accurate and complete scene geometry, rather than passively reconstructing scenes from pre-collected images. However, existing active reconstruction methods often rely on hand-crafted geometric heuristics, which can lead to redundant observations without substantially improving reconstruction quality. To address this limitation, we propose AREA3D, an active reconstruction agent that leverages feed-forward 3D reconstruction models and vision-language guidance. Our framework decouples view-uncertainty modeling from the underlying feed-forward reconstructor, enabling precise uncertainty estimation without expensive online optimization. In addition, an integrated vision-language model provides high-level semantic guidance, encouraging informative and diverse viewpoints beyond purely geometric cues. Extensive experiments on both scene-level and object-level benchmarks demonstrate that AREA3D achieves state-of-the-art reconstruction accuracy, particularly in the sparse-view regime. Code will be made available at: https://github.com/TianlingXu/AREA3D .

Paperid: 184, https://arxiv.org/pdf/2512.05119.pdf GitHub

Abstract:
In real-world scenarios, providing user queries with visually enhanced responses can considerably benefit understanding and memory, underscoring the great value of interleaved image-text generation. Despite recent progress, like the visual autoregressive model that unifies text and image processing in a single transformer architecture, generating high-quality interleaved content remains challenging. Moreover, evaluations of these interleaved sequences largely remain underexplored, with existing benchmarks often limited by unimodal metrics that inadequately assess the intricacies of combined image-text outputs. To address these issues, we present RAG-IGBench, a thorough benchmark designed specifically to evaluate the task of Interleaved Generation based on Retrieval-Augmented Generation (RAG-IG) in open-domain question answering. RAG-IG integrates multimodal large language models (MLLMs) with retrieval mechanisms, enabling the models to access external image-text information for generating coherent multimodal content. Distinct from previous datasets, RAG-IGBench draws on the latest publicly available content from social platforms and introduces innovative evaluation metrics that measure the quality of text and images, as well as their consistency. Through extensive experiments with state-of-the-art MLLMs (both open-source and proprietary) on RAG-IGBench, we provide an in-depth analysis examining the capabilities and limitations of these models. Additionally, we validate our evaluation metrics by demonstrating their high correlation with human assessments. Models fine-tuned on RAG-IGBench's training set exhibit improved performance across multiple benchmarks, confirming both the quality and practical utility of our dataset. Our benchmark is available at https://github.com/USTC-StarTeam/RAG-IGBench.

Paperid: 185, https://arxiv.org/pdf/2512.05113.pdf GitHub

Abstract:
Synthesizing high-fidelity frozen 3D scenes from monocular Mannequin-Challenge (MC) videos is a unique problem distinct from standard dynamic scene reconstruction. Instead of focusing on modeling motion, our goal is to create a frozen scene while strategically preserving subtle dynamics to enable user-controlled instant selection. To achieve this, we introduce a novel application of dynamic Gaussian splatting: the scene is modeled dynamically, which retains nearby temporal variation, and a static scene is rendered by fixing the model's time parameter. However, under this usage, monocular capture with sparse temporal supervision introduces artifacts like ghosting and blur for Gaussians that become unobserved or occluded at weakly supervised timestamps. We propose Splannequin, an architecture-agnostic regularization that detects two states of Gaussian primitives, hidden and defective, and applies temporal anchoring. Under predominantly forward camera motion, hidden states are anchored to their recent well-observed past states, while defective states are anchored to future states with stronger supervision. Our method integrates into existing dynamic Gaussian pipelines via simple loss terms, requires no architectural changes, and adds zero inference overhead. This results in markedly improved visual quality, enabling high-fidelity, user-selectable frozen-time renderings, validated by a 96% user preference. Project page: https://chien90190.github.io/splannequin/

Paperid: 186, https://arxiv.org/pdf/2512.05112.pdf GitHub

Abstract:
Recent unified multimodal large language models (MLLMs) have shown impressive capabilities, incorporating chain-of-thought (CoT) reasoning for enhanced text-to-image generation. However, existing approaches remain limited, either treating the model merely as a standalone generator or relying on abstract textual planning. To this end, we propose Draft-as-CoT (DraCo), a novel interleaved reasoning paradigm that fully leverages both textual and visual contents in CoT for better planning and verification. Our method first generates a low-resolution draft image as preview, providing more concrete and structural visual planning and guidance. Then, we employ the model's inherent understanding capability to verify potential semantic misalignments between the draft and input prompt, and performs refinement through selective corrections with super-resolution. In this way, our approach addresses two fundamental challenges: the coarse-grained nature of textual planning and the difficulty in generating rare attribute combinations. To support training, we curate DraCo-240K, aiming to enhance three atomic capabilities spanning general correction, instance manipulation, and layout reorganization. Supported by DraCo-CFG, a specialized classifier-free guidance (CFG) strategy for interleaved reasoning, DraCo achieves a tremendous increase on GenEval (+8%), Imagine-Bench (+0.91), and GenEval++ (+3%), significantly outperforming direct generation and other generation methods empowered by CoT.

Paperid: 187, https://arxiv.org/pdf/2512.05110.pdf GitHub

Abstract:
We introduce ShadowDraw, a framework that transforms ordinary 3D objects into shadow-drawing compositional art. Given a 3D object, our system predicts scene parameters, including object pose and lighting, together with a partial line drawing, such that the cast shadow completes the drawing into a recognizable image. To this end, we optimize scene configurations to reveal meaningful shadows, employ shadow strokes to guide line drawing generation, and adopt automatic evaluation to enforce shadow-drawing coherence and visual quality. Experiments show that ShadowDraw produces compelling results across diverse inputs, from real-world scans and curated datasets to generative assets, and naturally extends to multi-object scenes, animations, and physical deployments. Our work provides a practical pipeline for creating shadow-drawing art and broadens the design space of computational visual art, bridging the gap between algorithmic design and artistic storytelling. Check out our project page https://red-fairy.github.io/ShadowDraw/ for more results and an end-to-end real-world demonstration of our pipeline!

Paperid: 188, https://arxiv.org/pdf/2512.05106.pdf GitHub GitHub

Abstract:
Standard diffusion corrupts data using Gaussian noise whose Fourier coefficients have random magnitudes and random phases. While effective for unconditional or text-to-image generation, corrupting phase components destroys spatial structure, making it ill-suited for tasks requiring geometric consistency, such as re-rendering, simulation enhancement, and image-to-image translation. We introduce Phase-Preserving Diffusion ϕ-PD, a model-agnostic reformulation of the diffusion process that preserves input phase while randomizing magnitude, enabling structure-aligned generation without architectural changes or additional parameters. We further propose Frequency-Selective Structured (FSS) noise, which provides continuous control over structural rigidity via a single frequency-cutoff parameter. ϕ-PD adds no inference-time cost and is compatible with any diffusion model for images or videos. Across photorealistic and stylized re-rendering, as well as sim-to-real enhancement for driving planners, ϕ-PD produces controllable, spatially aligned results. When applied to the CARLA simulator, ϕ-PD improves CARLA-to-Waymo planner performance by 50\%. The method is complementary to existing conditioning approaches and broadly applicable to image-to-image and video-to-video generation. Videos, additional examples, and code are available on our \href{https://yuzeng-at-tri.github.io/ppd-page/}{project page}.

Paperid: 189, https://arxiv.org/pdf/2512.05105.pdf GitHub

Abstract:
Long context reasoning in large language models (LLMs) has demonstrated enhancement of their cognitive capabilities via chain-of-thought (CoT) inference. Training such models is usually done via reinforcement learning with verifiable rewards (RLVR) in reasoning based problems, like math and programming. However, RLVR is limited by several bottlenecks, such as, lack of dense reward, and inadequate sample efficiency. As a result, it requires significant compute resources in post-training phase. To overcome these limitations, in this work, we propose \textbf{Semantic Soft Bootstrapping (SSB)}, a self-distillation technique, in which the same base language model plays the role of both teacher and student, but receives different semantic contexts about the correctness of its outcome at training time. The model is first prompted with a math problem and several rollouts are generated. From them, the correct and most common incorrect response are filtered, and then provided to the model in context to produce a more robust, step-by-step explanation with a verified final answer. This pipeline automatically curates a paired teacher-student training set from raw problem-answer data, without any human intervention. This generation process also produces a sequence of logits, which is what the student model tries to match in the training phase just from the bare question alone. In our experiment, Qwen2.5-3B-Instruct on GSM8K dataset via parameter-efficient fine-tuning. We then tested its accuracy on MATH500, and AIME2024 benchmarks. Our experiments show a jump of 10.6%, and 10% improvements in accuracy, respectively, over group relative policy optimization (GRPO), which is a commonly used RLVR algorithm. Our code is available at https://github.com/purbeshmitra/semantic-soft-bootstrapping, and the model, curated dataset is available at https://huggingface.co/purbeshmitra/semantic-soft-bootstrapping.

Paperid: 190, https://arxiv.org/pdf/2512.05091.pdf GitHub

Abstract:
Recent advances in Multimodal Large Language Models (MLLMs) have significantly improved performance on tasks such as visual grounding and visual question answering. However, the reasoning processes of these models remain largely opaque; they typically output only final predictions without revealing the intermediate steps or fine-grained evidence (e.g., pixels, locations) that lead to the result. This contrasts with human intelligence, which naturally operates through a chain of visual reasoning. To address this limitation, we introduce the Visual Reasoning Tracer (VRT) task, which requires models to not only localize the target object but also explicitly predict the intermediate objects that form the reasoning path. To advance research in this area, we contribute: (1) VRT-Bench, a human-annotated benchmark for evaluating visual reasoning; (2) a new metric for assessing the quality of reasoning traces; and (3) VRT-80k, a large-scale dataset for reasoning model training. Our experiments reveal that while existing models often produce the correct final output, they struggle to ground their intermediate reasoning. In contrast, models trained on VRT-80k achieve substantial improvements in tracing the reasoning path.

Paperid: 191, https://arxiv.org/pdf/2512.05081.pdf GitHub

Abstract:
Recent advances in autoregressive video diffusion have enabled real-time frame streaming, yet existing solutions still suffer from temporal repetition, drift, and motion deceleration. We find that naively applying StreamingLLM-style attention sinks to video diffusion leads to fidelity degradation and motion stagnation. To overcome this, we introduce Deep Forcing, which consists of two training-free mechanisms that address this without any fine-tuning. Specifically, 1) Deep Sink dedicates half of the sliding window to persistent sink tokens and re-aligns their temporal RoPE phase to the current timeline, stabilizing global context during long rollouts. 2) Participative Compression performs importance-aware KV cache pruning that preserves only tokens actively participating in recent attention while safely discarding redundant and degraded history, minimizing error accumulation under out-of-distribution length generation. Together, these components enable over 12x extrapolation (e.g. 5s-trained to 60s+ generation) with better imaging quality than LongLive, better aesthetic quality than RollingForcing, almost maintaining overall consistency, and substantial gains in dynamic degree, all while maintaining real-time generation. Our results demonstrate that training-free KV-cache management can match or exceed training-based approaches for autoregressively streaming long-video generation.

Paperid: 192, https://arxiv.org/pdf/2512.05080.pdf GitHub

Abstract:
Structure-based drug design (SBDD) focuses on designing small-molecule ligands that bind to specific protein pockets. Computational methods are integral in modern SBDD workflows and often make use of virtual screening methods via docking or pharmacophore search. Modern generative modeling approaches have focused on improving novel ligand discovery by enabling de novo design. In this work, we recognize that these tasks share a common structure and can therefore be represented as different instantiations of a consistent generative modeling framework. We propose a unified approach in OMTRA, a multi-modal flow matching model that flexibly performs many tasks relevant to SBDD, including some with no analogue in conventional workflows. Additionally, we curate a dataset of 500M 3D molecular conformers, complementing protein-ligand data and expanding the chemical diversity available for training. OMTRA obtains state of the art performance on pocket-conditioned de novo design and docking; however, the effects of large-scale pretraining and multi-task training are modest. All code, trained models, and dataset for reproducing this work are available at https://github.com/gnina/OMTRA

Paperid: 193, https://arxiv.org/pdf/2512.05076.pdf GitHub

Abstract:
Emerging video diffusion models achieve high visual fidelity but fundamentally couple scene dynamics with camera motion, limiting their ability to provide precise spatial and temporal control. We introduce a 4D-controllable video diffusion framework that explicitly decouples scene dynamics from camera pose, enabling fine-grained manipulation of both scene dynamics and camera viewpoint. Our framework takes continuous world-time sequences and camera trajectories as conditioning inputs, injecting them into the video diffusion model through a 4D positional encoding in the attention layer and adaptive normalizations for feature modulation. To train this model, we curate a unique dataset in which temporal and camera variations are independently parameterized; this dataset will be made public. Experiments show that our model achieves robust real-world 4D control across diverse timing patterns and camera trajectories, while preserving high generation quality and outperforming prior work in controllability. See our website for video results: https://19reborn.github.io/Bullet4D/

Paperid: 194, https://arxiv.org/pdf/2512.05069.pdf GitHub

Abstract:
Unsupervised anomaly-based intrusion detection requires models that can generalize to attack patterns not observed during training. This work presents the first large-scale evaluation of hybrid quantum-classical (HQC) autoencoders for this task. We construct a unified experimental framework that iterates over key quantum design choices, including quantum-layer placement, measurement approach, variational and non-variational formulations, and latent-space regularization. Experiments across three benchmark NIDS datasets show that HQC autoencoders can match or exceed classical performance in their best configurations, although they exhibit higher sensitivity to architectural decisions. Under zero-day evaluation, well-configured HQC models provide stronger and more stable generalization than classical and supervised baselines. Simulated gate-noise experiments reveal early performance degradation, indicating the need for noise-aware HQC designs. These results provide the first data-driven characterization of HQC autoencoder behavior for network intrusion detection and outline key factors that govern their practical viability. All experiment code and configurations are available at https://github.com/arasyi/hqcae-network-intrusion-detection.

Paperid: 195, https://arxiv.org/pdf/2512.05060.pdf GitHub GitHub GitHub

Abstract:
Constructing 4D language fields is crucial for embodied AI, augmented/virtual reality, and 4D scene understanding, as they provide enriched semantic representations of dynamic environments and enable open-vocabulary querying in complex scenarios. However, existing approaches to 4D semantic field construction primarily rely on scene-specific Gaussian splatting, which requires per-scene optimization, exhibits limited generalization, and is difficult to scale to real-world applications. To address these limitations, we propose 4DLangVGGT, the first Transformer-based feed-forward unified framework for 4D language grounding, that jointly integrates geometric perception and language alignment within a single architecture. 4DLangVGGT has two key components: the 4D Visual Geometry Transformer, StreamVGGT, which captures spatio-temporal geometric representations of dynamic scenes; and the Semantic Bridging Decoder (SBD), which projects geometry-aware features into a language-aligned semantic space, thereby enhancing semantic interpretability while preserving structural fidelity. Unlike prior methods that depend on costly per-scene optimization, 4DLangVGGT can be jointly trained across multiple dynamic scenes and directly applied during inference, achieving both deployment efficiency and strong generalization. This design significantly improves the practicality of large-scale deployment and establishes a new paradigm for open-vocabulary 4D scene understanding. Experiments on HyperNeRF and Neu3D datasets demonstrate that our approach not only generalizes effectively but also achieves state-of-the-art performance, achieving up to 2% gains under per-scene training and 1% improvements under multi-scene training. Our code released in https://github.com/hustvl/4DLangVGGT

Paperid: 196, https://arxiv.org/pdf/2512.05044.pdf GitHub

Abstract:
Generating interactive and dynamic 4D scenes from a single static image remains a core challenge. Most existing generate-then-reconstruct and reconstruct-then-generate methods decouple geometry from motion, causing spatiotemporal inconsistencies and poor generalization. To address these, we extend the reconstruct-then-generate framework to jointly perform Motion generation and geometric Reconstruction for 4D Synthesis (MoRe4D). We first introduce TrajScene-60K, a large-scale dataset of 60,000 video samples with dense point trajectories, addressing the scarcity of high-quality 4D scene data. Based on this, we propose a diffusion-based 4D Scene Trajectory Generator (4D-STraG) to jointly generate geometrically consistent and motion-plausible 4D point trajectories. To leverage single-view priors, we design a depth-guided motion normalization strategy and a motion-aware module for effective geometry and dynamics integration. We then propose a 4D View Synthesis Module (4D-ViSM) to render videos with arbitrary camera trajectories from 4D point track representations. Experiments show that MoRe4D generates high-quality 4D scenes with multi-view consistency and rich dynamic details from a single image. Code: https://github.com/Zhangyr2022/MoRe4D.

Paperid: 197, https://arxiv.org/pdf/2512.05025.pdf GitHub

Abstract:
Earth observation (EO) data spans a wide range of spatial, spectral, and temporal resolutions, from high-resolution optical imagery to low resolution multispectral products or radar time series. While recent foundation models have improved multimodal integration for learning meaningful representations, they often expect fixed input resolutions or are based on sensor-specific encoders limiting generalization across heterogeneous EO modalities. To overcome these limitations we introduce RAMEN, a resolution-adjustable multimodal encoder that learns a shared visual representation across EO data in a fully sensor-agnostic manner. RAMEN treats the modality and spatial and temporal resolutions as key input data features, enabling coherent analysis across modalities within a unified latent space. Its main methodological contribution is to define spatial resolution as a controllable output parameter, giving users direct control over the desired level of detail at inference and allowing explicit trade-offs between spatial precision and computational cost. We train a single, unified transformer encoder reconstructing masked multimodal EO data drawn from diverse sources, ensuring generalization across sensors and resolutions. Once pretrained, RAMEN transfers effectively to both known and unseen sensor configurations and outperforms larger state-of-the-art models on the community-standard PANGAEA benchmark, containing various multi-sensor and multi-resolution downstream tasks. Our code and pretrained model are available at https://github.com/nicolashoudre/RAMEN.

Paperid: 198, https://arxiv.org/pdf/2512.04996.pdf GitHub

Abstract:
This paper proposes a memory-efficient optimization strategy for the high-performance point cloud registration algorithm VANICP, enabling lightweight execution on embedded GPUs with constrained hardware resources. VANICP is a recently published acceleration framework that significantly improves the computational efficiency of point-cloud-based applications. By transforming the global nearest neighbor search into a localized process through a dilation-based information propagation mechanism, VANICP greatly reduces the computational complexity of the NNS. However, its original implementation demands a considerable amount of memory, which restricts its deployment in resource-constrained environments such as embedded systems. To address this issue, we propose a GPU-oriented dynamic memory assignment strategy that optimizes the memory usage of the dilation operation. Furthermore, based on this strategy, we construct an enhanced version of the VANICP framework that achieves over 97% reduction in memory consumption while preserving the original performance. Source code is published on: https://github.com/changqiong/VANICP4Em.git.

Paperid: 199, https://arxiv.org/pdf/2512.04969.pdf GitHub

Abstract:
Rich feature representations derived from CLIP-ViT have been widely utilized in AI-generated image detection. While most existing methods primarily leverage features from the final layer, we systematically analyze the contributions of layer-wise features to this task. Our study reveals that earlier layers provide more localized and generalizable features, often surpassing the performance of final-layer features in detection tasks. Moreover, we find that different layers capture distinct aspects of the data, each contributing uniquely to AI-generated image detection. Motivated by these findings, we introduce a novel adaptive method, termed MoLD, which dynamically integrates features from multiple ViT layers using a gating-based mechanism. Extensive experiments on both GAN- and diffusion-generated images demonstrate that MoLD significantly improves detection performance, enhances generalization across diverse generative models, and exhibits robustness in real-world scenarios. Finally, we illustrate the scalability and versatility of our approach by successfully applying it to other pre-trained ViTs, such as DINOv2.

Paperid: 200, https://arxiv.org/pdf/2512.04939.pdf GitHub

Abstract:
3D vision foundation models like Visual Geometry Grounded Transformer (VGGT) have advanced greatly in geometric perception. However, it is time-consuming and memory-intensive for long sequences, limiting application to large-scale scenes beyond hundreds of images. To address this, we propose LiteVGGT, achieving up to 10x speedup and substantial memory reduction, enabling efficient processing of 1000-image scenes. We derive two key insights for 3D reconstruction: (1) tokens from local image regions have inherent geometric correlations, leading to high similarity and computational redundancy; (2) token similarity across adjacent network layers remains stable, allowing for reusable merge decisions. Guided by these, we design a simple yet efficient strategy, dubbed geometry-aware cached token merging. We analyze each token's geometric importance, optimizing anchor token selection to better preserve key information for reconstruction. We also cache and reuse merge indices across layers, substantially reducing latency with minimal accuracy impact. This strategy retains VGGT's core performance, enabling efficient fine-tuning and FP8 quantization for further gains. Extensive experiments validate LiteVGGT's effectiveness, scalability, and robustness. Project page: https://garlicba.github.io/LiteVGGT/

Paperid: 201, https://arxiv.org/pdf/2512.04926.pdf GitHub

Abstract:
Latent Diffusion Models (LDMs) inherently follow a coarse-to-fine generation process, where high-level semantic structure is generated slightly earlier than fine-grained texture. This indicates the preceding semantics potentially benefit texture generation by providing a semantic anchor. Recent advances have integrated semantic priors from pretrained visual encoders to further enhance LDMs, yet they still denoise semantic and VAE-encoded texture synchronously, neglecting such ordering. Observing these, we propose Semantic-First Diffusion (SFD), a latent diffusion paradigm that explicitly prioritizes semantic formation. SFD first constructs composite latents by combining a compact semantic latent, which is extracted from a pretrained visual encoder via a dedicated Semantic VAE, with the texture latent. The core of SFD is to denoise the semantic and texture latents asynchronously using separate noise schedules: semantics precede textures by a temporal offset, providing clearer high-level guidance for texture refinement and enabling natural coarse-to-fine generation. On ImageNet 256x256 with guidance, SFD achieves FID 1.06 (LightningDiT-XL) and FID 1.04 (1.0B LightningDiT-XXL), while achieving up to 100x faster convergence than the original DiT. SFD also improves existing methods like ReDi and VA-VAE, demonstrating the effectiveness of asynchronous, semantics-led modeling. Project page and code: https://yuemingpan.github.io/SFD.github.io/.

Paperid: 202, https://arxiv.org/pdf/2512.04883.pdf GitHub

Abstract:
Real-time tracking of small unmanned aerial vehicles (UAVs) on edge devices faces a fundamental resolution-speed conflict. Downsampling high-resolution imagery to standard detector input sizes causes small target features to collapse below detectable thresholds. Yet processing native 1080p frames on resource-constrained platforms yields insufficient throughput for smooth gimbal control. We propose SDG-Track, a Sparse Detection-Guided Tracker that adopts an Observer-Follower architecture to reconcile this conflict. The Observer stream runs a high-capacity detector at low frequency on the GPU to provide accurate position anchors from 1920x1080 frames. The Follower stream performs high-frequency trajectory interpolation via ROI-constrained sparse optical flow on the CPU. To handle tracking failures from occlusion or model drift caused by spectrally similar distractors, we introduce Dual-Space Recovery, a training-free re-acquisition mechanism combining color histogram matching with geometric consistency constraints. Experiments on a ground-to-air tracking station demonstrate that SDG-Track achieves 35.1 FPS system throughput while retaining 97.2\% of the frame-by-frame detection precision. The system successfully tracks agile FPV drones under real-world operational conditions on an NVIDIA Jetson Orin Nano. Our paper code is publicly available at https://github.com/Jeffry-wen/SDG-Track

Paperid: 203, https://arxiv.org/pdf/2512.04841.pdf GitHub

Abstract:
Large Language Models (LLMs) exhibit remarkable capabilities but remain vulnerable to adversarial manipulations such as jailbreaking, where crafted prompts bypass safety mechanisms. Understanding the causal factors behind such vulnerabilities is essential for building reliable defenses. In this work, we introduce a unified causality analysis framework that systematically supports all levels of causal investigation in LLMs, ranging from token-level, neuron-level, and layer-level interventions to representation-level analysis. The framework enables consistent experimentation and comparison across diverse causality-based attack and defense methods. Accompanying this implementation, we provide the first comprehensive survey of causality-driven jailbreak studies and empirically evaluate the framework on multiple open-weight models and safety-critical benchmarks including jailbreaks, hallucination detection, backdoor identification, and fairness evaluation. Our results reveal that: (1) targeted interventions on causally critical components can reliably modify safety behavior; (2) safety-related mechanisms are highly localized (i.e., concentrated in early-to-middle layers with only 1--2\% of neurons exhibiting causal influence); and (3) causal features extracted from our framework achieve over 95\% detection accuracy across multiple threat types. By bridging theoretical causality analysis and practical model safety, our framework establishes a reproducible foundation for research on causality-based attacks, interpretability, and robust attack detection and mitigation in LLMs. Code is available at https://github.com/Amadeuszhao/SOK_Casuality.

Paperid: 204, https://arxiv.org/pdf/2512.04813.pdf GitHub

Abstract:
Imitation learning method has shown immense promise for robotic manipulation, yet its practical deployment is fundamentally constrained by the data scarcity. Despite prior work on collecting large-scale datasets, there still remains a significant gap to robust spatial generalization. We identify a key limitation: individual trajectories, regardless of their length, are typically collected from a \emph{single, static spatial configuration} of the environment. This includes fixed object and target spatial positions as well as unchanging camera viewpoints, which significantly restricts the diversity of spatial information available for learning. To address this critical bottleneck in data efficiency, we propose \textbf{MOtion-Based Variability Enhancement} (\emph{MOVE}), a simple yet effective data collection paradigm that enables the acquisition of richer spatial information from dynamic demonstrations. Our core contribution is an augmentation strategy that injects motion into any movable objects within the environment for each demonstration. This process implicitly generates a dense and diverse set of spatial configurations within a single trajectory. We conduct extensive experiments in both simulation and real-world environments to validate our approach. For example, in simulation tasks requiring strong spatial generalization, \emph{MOVE} achieves an average success rate of 39.1\%, a 76.1\% relative improvement over the static data collection paradigm (22.2\%), and yields up to 2--5$\times$ gains in data efficiency on certain tasks. Our code is available at https://github.com/lucywang720/MOVE.

Paperid: 205, https://arxiv.org/pdf/2512.04810.pdf GitHub

Abstract:
We propose EMMA, an efficient and unified architecture for multimodal understanding, generation and editing. Specifically, EMMA primarily consists of 1) An efficient autoencoder with a 32x compression ratio, which significantly reduces the number of tokens required for generation. This also ensures the training balance between understanding and generation tasks by applying the same compression ratio to images. 2) Channel-wise concatenation instead of token-wise concatenation among visual understanding and generation tokens, which further reduces the visual tokens in unified architectures. 3) A shared-and-decoupled network that enables mutual improvements across tasks while meeting the task-specific modeling requirements. 4) A mixture-of-experts mechanism adopted for visual understanding encoder, which substantially improves perceptual capabilities with a few parameters increase. Extensive experiments have shown that EMMA-4B can significantly outperform state-of-the-art unified multimodal approaches (e.g., BAGEL-7B) in both efficiency and performance, while also achieving competitive results compared to recent multimodal understanding and generation experts (e.g., Qwen3-VL and Qwen-Image). We believe that EMMA lays a solid foundation for the future development of unified multimodal architectures.

Paperid: 206, https://arxiv.org/pdf/2512.04786.pdf GitHub

Abstract:
Generating high-fidelity, seamless textures directly on 3D surfaces, what we term 3D-native texturing, remains a fundamental open challenge, with the potential to overcome long-standing limitations of UV-based and multi-view projection methods. However, existing native approaches are constrained by the absence of a powerful and versatile latent representation, which severely limits the fidelity and generality of their generated textures. We identify this representation gap as the principal barrier to further progress. We introduce LaFiTe, a framework that addresses this challenge by learning to generate textures as a 3D generative sparse latent color field. At its core, LaFiTe employs a variational autoencoder (VAE) to encode complex surface appearance into a sparse, structured latent space, which is subsequently decoded into a continuous color field. This representation achieves unprecedented fidelity, exceeding state-of-the-art methods by >10 dB PSNR in reconstruction, by effectively disentangling texture appearance from mesh topology and UV parameterization. Building upon this strong representation, a conditional rectified-flow model synthesizes high-quality, coherent textures across diverse styles and geometries. Extensive experiments demonstrate that LaFiTe not only sets a new benchmark for 3D-native texturing but also enables flexible downstream applications such as material synthesis and texture super-resolution, paving the way for the next generation of 3D content creation workflows.

Paperid: 207, https://arxiv.org/pdf/2512.04784.pdf GitHub

Abstract:
Consistent image generation requires faithfully preserving identities, styles, and logical coherence across multiple images, which is essential for applications such as storytelling and character design. Supervised training approaches struggle with this task due to the lack of large-scale datasets capturing visual consistency and the complexity of modeling human perceptual preferences. In this paper, we argue that reinforcement learning (RL) offers a promising alternative by enabling models to learn complex and subjective visual criteria in a data-free manner. To achieve this, we introduce PaCo-RL, a comprehensive framework that combines a specialized consistency reward model with an efficient RL algorithm. The first component, PaCo-Reward, is a pairwise consistency evaluator trained on a large-scale dataset constructed via automated sub-figure pairing. It evaluates consistency through a generative, autoregressive scoring mechanism enhanced by task-aware instructions and CoT reasons. The second component, PaCo-GRPO, leverages a novel resolution-decoupled optimization strategy to substantially reduce RL cost, alongside a log-tamed multi-reward aggregation mechanism that ensures balanced and stable reward optimization. Extensive experiments across the two representative subtasks show that PaCo-Reward significantly improves alignment with human perceptions of visual consistency, and PaCo-GRPO achieves state-of-the-art consistency performance with improved training efficiency and stability. Together, these results highlight the promise of PaCo-RL as a practical and scalable solution for consistent image generation. The project page is available at https://x-gengroup.github.io/HomePage_PaCo-RL/.

Paperid: 208, https://arxiv.org/pdf/2512.04770.pdf GitHub

Abstract:
Brain-body co-evolution enables animals to develop complex behaviors in their environments. Inspired by this biological synergy, embodied co-design (ECD) has emerged as a transformative paradigm for creating intelligent agents-from virtual creatures to physical robots-by jointly optimizing their morphologies and controllers rather than treating control in isolation. This integrated approach facilitates richer environmental interactions and robust task performance. In this survey, we provide a systematic overview of recent advances in ECD. We first formalize the concept of ECD and position it within related fields. We then introduce a hierarchical taxonomy: a lower layer that breaks down agent design into three fundamental components-controlling brain, body morphology, and task environment-and an upper layer that integrates these components into four major ECD frameworks: bi-level, single-level, generative, and open-ended. This taxonomy allows us to synthesize insights from more than one hundred recent studies. We further review notable benchmarks, datasets, and applications in both simulated and real-world scenarios. Finally, we identify significant challenges and offer insights into promising future research directions. A project associated with this survey has been created at https://github.com/Yuxing-Wang-THU/SurveyBrainBody.

Paperid: 209, https://arxiv.org/pdf/2512.04761.pdf GitHub

Abstract:
VR sketching lets users explore and iterate on ideas directly in 3D, offering a faster and more intuitive alternative to conventional CAD tools. However, existing sketch-to-shape models ignore the temporal ordering of strokes, discarding crucial cues about structure and design intent. We introduce VRSketch2Shape, the first framework and multi-category dataset for generating 3D shapes from sequential VR sketches. Our contributions are threefold: (i) an automated pipeline that generates sequential VR sketches from arbitrary shapes, (ii) a dataset of over 20k synthetic and 900 hand-drawn sketch-shape pairs across four categories, and (iii) an order-aware sketch encoder coupled with a diffusion-based 3D generator. Our approach yields higher geometric fidelity than prior work, generalizes effectively from synthetic to real sketches with minimal supervision, and performs well even on partial sketches. All data and models will be released open-source at https://chenyizi086.github.io/VRSketch2Shape_website.

Paperid: 210, https://arxiv.org/pdf/2512.04746.pdf GitHub

Abstract:
Extreme low-bit quantization is critical for efficiently deploying Large Language Models (LLMs), yet it often leads to severe performance degradation at 2-bits and even 4-bits (e.g., MXFP4). We present SignRoundV2, a post-training quantization framework that is highly effective even without mixed-precision. SignRoundV2 introduces (1) a fast sensitivity metric that combines gradient information with quantization-induced deviations to guide layer-wise bit allocation, and (2) a lightweight pre-tuning search for quantization scales to improve extremely low-bit quantization. These components allow SignRoundV2 to close the gap with full-precision models. Extensive experiments indicate that our method sustains competitive accuracy for LLMs, achieving production-grade performance with about 1 percent variance at 4-5 bits and strong results even at 2 bits. The implementation is available at https://github.com/intel/auto-round.

Paperid: 211, https://arxiv.org/pdf/2512.04720.pdf GitHub

Abstract:
Non-autoregressive (NAR) text-to-speech synthesis relies on length alignment between text sequences and audio representations, constraining naturalness and expressiveness. Existing methods depend on duration modeling or pseudo-alignment strategies that severely limit naturalness and computational efficiency. We propose M3-TTS, a concise and efficient NAR TTS paradigm based on multi-modal diffusion transformer (MM-DiT) architecture. M3-TTS employs joint diffusion transformer layers for cross-modal alignment, achieving stable monotonic alignment between variable-length text-speech sequences without pseudo-alignment requirements. Single diffusion transformer layers further enhance acoustic detail modeling. The framework integrates a mel-vae codec that provides 3* training acceleration. Experimental results on Seed-TTS and AISHELL-3 benchmarks demonstrate that M3-TTS achieves state-of-the-art NAR performance with the lowest word error rates (1.36\% English, 1.31\% Chinese) while maintaining competitive naturalness scores. Code and demos will be available at https://wwwwxp.github.io/M3-TTS.

Paperid: 212, https://arxiv.org/pdf/2512.04699.pdf GitHub

Abstract:
Arbitrary-scale super-resolution (ASSR) overcomes the limitation of traditional super-resolution (SR) methods that operate only at fixed scales (e.g., 4x), enabling a single model to handle arbitrary magnification. Most existing ASSR approaches rely on implicit neural representation (INR), but its regression-driven feature extraction and aggregation intrinsically limit the ability to synthesize fine details, leading to low realism. Recent diffusion-based realistic image super-resolution (Real-ISR) models leverage powerful pre-trained diffusion priors and show impressive results at the 4x setting. We observe that they can also achieve ASSR because the diffusion prior implicitly adapts to scale by encouraging high-realism generation. However, without explicit scale control, the diffusion process cannot be properly adjusted for different magnification levels, resulting in excessive hallucination or blurry outputs, especially under ultra-high scales. To address these issues, we propose OmniScaleSR, a diffusion-based realistic arbitrary-scale SR framework designed to achieve both high fidelity and high realism. We introduce explicit, diffusion-native scale control mechanisms that work synergistically with implicit scale adaptation, enabling scale-aware and content-aware modulation of the diffusion process. In addition, we incorporate multi-domain fidelity enhancement designs to further improve reconstruction accuracy. Extensive experiments on bicubic degradation benchmarks and real-world datasets show that OmniScaleSR surpasses state-of-the-art methods in both fidelity and perceptual realism, with particularly strong performance at large magnification factors. Code will be released at https://github.com/chaixinning/OmniScaleSR.

Paperid: 213, https://arxiv.org/pdf/2512.04694.pdf GitHub

Abstract:
Effective earthquake risk reduction relies on accurate site-specific evaluations. This requires models that can represent the influence of local site conditions on ground motion characteristics. In this context, data driven approaches that learn site controlled signatures from recorded ground motions offer a promising direction. We address strong ground motion generation from time-domain accelerometer records and introduce the TimesNet-Gen, a time-domain conditional generator. The approach uses a station specific latent bottleneck. We evaluate generation by comparing HVSR curves and fundamental site-frequency $f_0$ distributions between real and generated records per station, and summarize station specificity with a score based on the $f_0$ distribution confusion matrices. TimesNet-Gen achieves strong station-wise alignment and compares favorably with a spectrogram-based conditional VAE baseline for site-specific strong motion synthesis. Our codes are available via https://github.com/brsylmz23/TimesNet-Gen.

Paperid: 214, https://arxiv.org/pdf/2512.04686.pdf GitHub

Abstract:
Cross-view correspondence is a fundamental capability for spatial understanding and embodied AI. However, it is still far from being realized in Vision-Language Models (VLMs), especially in achieving precise point-level correspondence, which is crucial for precise affordance interaction. So we propose the Cross-View Point Correspondence (CVPC) task and CrossPoint-Bench, a comprehensive benchmark with hierarchical design, inspired by the human cognitive process of "perceive", "reason", and "correspond". Our evaluation shows the state-of-the-art models (e.g., Gemini-2.5-Pro) still fall far behind humans, with a gap of over 54.65% in overall accuracy, exposing a challenge in transitioning from coarse-grained judgement to fine-grained coordinate prediction. To address this problem, we construct CrossPoint-378K, a dataset with 378K question-answering pairs across 900 scenes, focused on actionable affordance regions that better reflect real-world manipulation and interaction scenarios. Furthermore, we propose CroPond that trained on the CrossPoint-378K dataset. Our CroPond achieves state-of-the-art performance on CrossPoint-Bench, surpassing Gemini-2.5-Pro by 39.7% accuracy, which offers a foundation for advancing future work on cross-view correspondence. The benchmark, dataset, and model are publicly available at https://github.com/WangYipu2002/CrossPoint.

Paperid: 215, https://arxiv.org/pdf/2512.04625.pdf GitHub

Abstract:
In the history of knowledge distillation, the focus has once shifted over time from logit-based to feature-based approaches. However, this transition has been revisited with the advent of Decoupled Knowledge Distillation (DKD), which re-emphasizes the importance of logit knowledge through advanced decoupling and weighting strategies. While DKD marks a significant advancement, its underlying mechanisms merit deeper exploration. As a response, we rethink DKD from a predictive distribution perspective. First, we introduce an enhanced version, the Generalized Decoupled Knowledge Distillation (GDKD) loss, which offers a more versatile method for decoupling logits. Then we pay particular attention to the teacher model's predictive distribution and its impact on the gradients of GDKD loss, uncovering two critical insights often overlooked: (1) the partitioning by the top logit considerably improves the interrelationship of non-top logits, and (2) amplifying the focus on the distillation loss of non-top logits enhances the knowledge extraction among them. Utilizing these insights, we further propose a streamlined GDKD algorithm with an efficient partition strategy to handle the multimodality of teacher models' predictive distribution. Our comprehensive experiments conducted on a variety of benchmarks, including CIFAR-100, ImageNet, Tiny-ImageNet, CUB-200-2011, and Cityscapes, demonstrate GDKD's superior performance over both the original DKD and other leading knowledge distillation methods. The code is available at https://github.com/ZaberKo/GDKD.

Paperid: 216, https://arxiv.org/pdf/2512.04578.pdf GitHub

Abstract:
Legal general intelligence (GI) refers to artificial intelligence (AI) that encompasses legal understanding, reasoning, and decision-making, simulating the expertise of legal experts across domains. However, existing benchmarks are result-oriented and fail to systematically evaluate the legal intelligence of large language models (LLMs), hindering the development of legal GI. To address this, we propose LexGenius, an expert-level Chinese legal benchmark for evaluating legal GI in LLMs. It follows a Dimension-Task-Ability framework, covering seven dimensions, eleven tasks, and twenty abilities. We use the recent legal cases and exam questions to create multiple-choice questions with a combination of manual and LLM reviews to reduce data leakage risks, ensuring accuracy and reliability through multiple rounds of checks. We evaluate 12 state-of-the-art LLMs using LexGenius and conduct an in-depth analysis. We find significant disparities across legal intelligence abilities for LLMs, with even the best LLMs lagging behind human legal professionals. We believe LexGenius can assess the legal intelligence abilities of LLMs and enhance legal GI development. Our project is available at https://github.com/QwenQKing/LexGenius.

Paperid: 217, https://arxiv.org/pdf/2512.04552.pdf GitHub

Abstract:
Differentiable reinforcement learning (RL) frameworks like DiffRO offer a powerful approach for controllable text-to-speech (TTS), but are vulnerable to reward hacking, particularly for nuanced tasks like emotion control. The policy model can exploit a vanilla Reward Model (RM) by generating acoustic artifacts to achieve spurious rewards, but at the cost of degrading perceptual quality. To address this, we propose Robust Reward Policy Optimization (RRPO), a novel framework that employs a hybrid regularization scheme. This scheme develops a robust RM whose reward signal is more reliably aligned with human perception, compelling the policy to abandon detrimental shortcuts and instead learn the complex features of genuine emotions. Our ablation study confirms the enhanced robustness of our RM, as evidenced by its strong cross-lingual generalization. The subjective evaluation demonstrates that this robust RM effectively mitigates reward hacking, leading to significant improvements in both emotional expressiveness and naturalness over all baselines. Demo page: https://lrwinr.github.io/RRPO-CosyVoice.

Paperid: 218, https://arxiv.org/pdf/2512.04534.pdf GitHub

Abstract:
Recent advances in diffusion models have brought remarkable progress in image and video editing, yet some tasks remain underexplored. In this paper, we introduce a new task, Object Retexture, which transfers local textures from a reference object to a target object in images or videos. To perform this task, a straightforward solution is to use ControlNet conditioned on the source structure and the reference texture. However, this approach suffers from limited controllability for two reasons: conditioning on the raw reference image introduces unwanted structural information, and it fails to disentangle the visual texture and structure information of the source. To address this problem, we propose Refaçade, a method that consists of two key designs to achieve precise and controllable texture transfer in both images and videos. First, we employ a texture remover trained on paired textured/untextured 3D mesh renderings to remove appearance information while preserving the geometry and motion of source videos. Second, we disrupt the reference global layout using a jigsaw permutation, encouraging the model to focus on local texture statistics rather than the global layout of the object. Extensive experiments demonstrate superior visual quality, precise editing, and controllability, outperforming strong baselines in both quantitative and human evaluations. Code is available at https://github.com/fishZe233/Refacade.

Paperid: 219, https://arxiv.org/pdf/2512.04528.pdf GitHub

Abstract:
Traditional high-quality 3D scanning and reconstruction typically relies on human labor to plan the scanning procedure. With the rapid development of embodied systems such as drones and robots, there is a growing demand of performing accurate 3D scanning and reconstruction in an fully automated manner. We introduce Auto3R, a data-driven uncertainty quantification model that is designed to automate the 3D scanning and reconstruction of scenes and objects, including objects with non-lambertian and specular materials. Specifically, in a process of iterative 3D reconstruction and scanning, Auto3R can make efficient and accurate prediction of uncertainty distribution over potential scanning viewpoints, without knowing the ground truth geometry and appearance. Through extensive experiments, Auto3R achieves superior performance that outperforms the state-of-the-art methods by a large margin. We also deploy Auto3R on a robot arm equipped with a camera and demonstrate that Auto3R can be used to effectively digitize real-world 3D objects and delivers ready-to-use and photorealistic digital assets. Our homepage: https://tomatoma00.github.io/auto3r.github.io .

Paperid: 220, https://arxiv.org/pdf/2512.04520.pdf GitHub

Abstract:
Due to the scarcity of annotated data and the substantial computational costs of model, conventional tuning methods in medical image segmentation face critical challenges. Current approaches to adapting pretrained models, including full-parameter and parameter-efficient fine-tuning, still rely heavily on task-specific training on downstream tasks. Therefore, zero-shot segmentation has gained increasing attention, especially with foundation models such as SAM demonstrating promising generalization capabilities. However, SAM still faces notable limitations on medical datasets due to domain shifts, making efficient zero-shot enhancement an urgent research goal. To address these challenges, we propose BA-TTA-SAM, a task-agnostic test-time adaptation framework that significantly enhances the zero-shot segmentation performance of SAM via test-time adaptation. This framework integrates two key mechanisms: (1) The encoder-level Gaussian prompt injection embeds Gaussian-based prompts directly into the image encoder, providing explicit guidance for initial representation learning. (2) The cross-layer boundary-aware attention alignment exploits the hierarchical feature interactions within the ViT backbone, aligning deep semantic responses with shallow boundary cues. Experiments on four datasets, including ISIC, Kvasir, BUSI, and REFUGE, show an average improvement of 12.4\% in the DICE score compared with SAM's zero-shot segmentation performance. The results demonstrate that our method consistently outperforms state-of-the-art models in medical image segmentation. Our framework significantly enhances the generalization ability of SAM, without requiring any source-domain training data. Extensive experiments on publicly available medical datasets strongly demonstrate the superiority of our framework. Our code is available at https://github.com/Emilychenlin/BA-TTA-SAM.

Paperid: 221, https://arxiv.org/pdf/2512.04515.pdf GitHub GitHub

Abstract:
Generating long, coherent egocentric videos is difficult, as hand-object interactions and procedural tasks require reliable long-term memory. Existing autoregressive models suffer from content drift, where object identity and scene semantics degrade over time. To address this challenge, we introduce EgoLCD, an end-to-end framework for egocentric long-context video generation that treats long video synthesis as a problem of efficient and stable memory management. EgoLCD combines a Long-Term Sparse KV Cache for stable global context with an attention-based short-term memory, extended by LoRA for local adaptation. A Memory Regulation Loss enforces consistent memory usage, and Structured Narrative Prompting provides explicit temporal guidance. Extensive experiments on the EgoVid-5M benchmark demonstrate that EgoLCD achieves state-of-the-art performance in both perceptual quality and temporal consistency, effectively mitigating generative forgetting and representing a significant step toward building scalable world models for embodied AI. Code: https://github.com/AIGeeksGroup/EgoLCD. Website: https://aigeeksgroup.github.io/EgoLCD.

Paperid: 222, https://arxiv.org/pdf/2512.04504.pdf GitHub GitHub GitHub

Abstract:
Recent image diffusion transformers achieve high-fidelity generation, but struggle to generate images beyond these scales, suffering from content repetition and quality degradation. In this work, we present UltraImage, a principled framework that addresses both issues. Through frequency-wise analysis of positional embeddings, we identify that repetition arises from the periodicity of the dominant frequency, whose period aligns with the training resolution. We introduce a recursive dominant frequency correction to constrain it within a single period after extrapolation. Furthermore, we find that quality degradation stems from diluted attention and thus propose entropy-guided adaptive attention concentration, which assigns higher focus factors to sharpen local attention for fine detail and lower ones to global attention patterns to preserve structural consistency. Experiments show that UltraImage consistently outperforms prior methods on Qwen-Image and Flux (around 4K) across three generation scenarios, reducing repetition and improving visual fidelity. Moreover, UltraImage can generate images up to 6K*6K without low-resolution guidance from a training resolution of 1328p, demonstrating its extreme extrapolation capability. Project page is available at \href{https://thu-ml.github.io/ultraimage.github.io/}{https://thu-ml.github.io/ultraimage.github.io/}.

Paperid: 223, https://arxiv.org/pdf/2512.04496.pdf GitHub

Abstract:
Inevitable specular highlights in practical environments severely impair the visual performance, thus degrading the task effectiveness and efficiency. Although there exist considerable methods that focus on local information from convolutional neural network models or global information from transformer models, the single-type model falls into a modeling dilemma between local fine-grained details and global long-range dependencies, thus deteriorating for specular highlights with different scales. Therefore, to accommodate specular highlights of all scales, we propose a multi-model architecture for specular highlight removal (MM-SHR) that effectively captures fine-grained features in highlight regions and models long-range dependencies between highlight and highlight-free areas. Specifically, we employ convolution operations to extract local details in the shallow layers of MM-SHR, and utilize the attention mechanism to capture global features in the deep layers, ensuring both operation efficiency and removal accuracy. To model long-range dependencies without compromising computational complexity, we utilize a coarse-to-fine manner and propose Omni-Directional Attention Integration Block(OAIBlock) and Adaptive Region-Aware Hybrid-Domain Dual Attention Convolutional Network(HDDAConv) , which leverage omni-directiona pixel-shifting and window-dividing operations at the raw features to achieve specular highlight removal. Extensive experimental results on three benchmark tasks and six types of surface materials demonstrate that MM-SHR outperforms state-of-the-art methods in both accuracy and efficiency for specular highlight removal. The implementation will be made publicly available at https://github.com/Htcicv/MM-SHR.

Paperid: 224, https://arxiv.org/pdf/2512.04463.pdf GitHub

Abstract:
We present a comparative study of multi-agent reinforcement learning (MARL) algorithms for cooperative warehouse robotics. We evaluate QMIX and IPPO on the Robotic Warehouse (RWARE) environment and a custom Unity 3D simulation. Our experiments reveal that QMIX's value decomposition significantly outperforms independent learning approaches (achieving 3.25 mean return vs. 0.38 for advanced IPPO), but requires extensive hyperparameter tuning -- particularly extended epsilon annealing (5M+ steps) for sparse reward discovery. We demonstrate successful deployment in Unity ML-Agents, achieving consistent package delivery after 1M training steps. While MARL shows promise for small-scale deployments (2-4 robots), significant scaling challenges remain. Code and analyses: https://pallman14.github.io/MARL-QMIX-Warehouse-Robots/

Paperid: 225, https://arxiv.org/pdf/2512.04457.pdf GitHub

Abstract:
Removing specific data influence from large language models (LLMs) remains challenging, as retraining is costly and existing approximate unlearning methods are often unstable. The challenge is exacerbated when the forget set is small or imbalanced. We introduce RapidUn, an influence-driven and parameter-efficient unlearning framework. It first estimates per-sample influence through a fast estimation module, then maps these scores into adaptive update weights that guide selective parameter updates -- forgetting harmful behavior while retaining general knowledge. On Mistral-7B and Llama-3-8B across Dolly-15k and Alpaca-57k, RapidUn achieves up to 100 times higher efficiency than full retraining and consistently outperforms Fisher, GA, and LoReUn on both in-distribution and out-of-distribution forgetting. These results establish influence-guided parameter reweighting as a scalable and interpretable paradigm for LLM unlearning.

Paperid: 226, https://arxiv.org/pdf/2512.04456.pdf GitHub

Abstract:
Recent image denoising methods have leveraged generative modeling for real noise synthesis to address the costly acquisition of real-world noisy data. However, these generative models typically require camera metadata and extensive target-specific noisy-clean image pairs, often showing limited generalization between settings. In this paper, to mitigate the prerequisites, we propose a Single-Pair Guided Diffusion for generalized noise synthesis GuidNoise, which uses a single noisy/clean pair as the guidance, often easily obtained by itself within a training set. To train GuidNoise, which generates synthetic noisy images from the guidance, we introduce a guidance-aware affine feature modification (GAFM) and a noise-aware refine loss to leverage the inherent potential of diffusion models. This loss function refines the diffusion model's backward process, making the model more adept at generating realistic noise distributions. The GuidNoise synthesizes high-quality noisy images under diverse noise environments without additional metadata during both training and inference. Additionally, GuidNoise enables the efficient generation of noisy-clean image pairs at inference time, making synthetic noise readily applicable for augmenting training data. This self-augmentation significantly improves denoising performance, especially in practical scenarios with lightweight models and limited training data. The code is available at https://github.com/chjinny/GuidNoise.

Paperid: 227, https://arxiv.org/pdf/2512.04445.pdf GitHub

Abstract:
Workflow automation promises substantial productivity gains in everyday document-related tasks. While prior agentic systems can execute isolated instructions, they struggle with automating multi-step, session-level workflows due to limited control over the operational process. To this end, we introduce AutoDW, a novel execution framework that enables stepwise, rollback-enabled operation orchestration. AutoDW incrementally plans API actions conditioned on user instructions, intent-filtered API candidates, and the evolving states of the document. It further employs robust rollback mechanisms at both the argument and API levels, enabling dynamic correction and fault tolerance. These designs together ensure that the execution trajectory of AutoDW remains aligned with user intent and document context across long-horizon workflows. To assess its effectiveness, we construct a comprehensive benchmark of 250 sessions and 1,708 human-annotated instructions, reflecting realistic document processing scenarios with interdependent instructions. AutoDW achieves 90% and 62% completion rates on instruction- and session-level tasks, respectively, outperforming strong baselines by 40% and 76%. Moreover, AutoDW also remains robust for the decision of backbone LLMs and on tasks with varying difficulty. Code and data will be open-sourced. Code: https://github.com/YJett/AutoDW

Paperid: 228, https://arxiv.org/pdf/2512.04434.pdf GitHub

Abstract:
Fast, geometry-generalizing surrogates for unsteady flow remain challenging. We present a time-dependent, geometry-aware Deep Operator Network that predicts velocity fields for moderate-Re flows around parametric and non-parametric shapes. The model encodes geometry via a signed distance field (SDF) trunk and flow history via a CNN branch, trained on 841 high-fidelity simulations. On held-out shapes, it attains $\sim 5\%$ relative L2 single-step error and up to 1000X speedups over CFD. We provide physics-centric rollout diagnostics, including phase error at probes and divergence norms, to quantify long-horizon fidelity. These reveal accurate near-term transients but error accumulation in fine-scale wakes, most pronounced for sharp-cornered geometries. We analyze failure modes and outline practical mitigations. Code, splits, and scripts are openly released at: https://github.com/baskargroup/TimeDependent-DeepONet to support reproducibility and benchmarking.

Paperid: 229, https://arxiv.org/pdf/2512.04426.pdf GitHub

Abstract:
As a challenging video editing task, movie trailer generation involves selecting and reorganizing movie shots to create engaging trailers. Currently, most existing automatic trailer generation methods employ a "selection-then-ranking" paradigm (i.e., first selecting key shots and then ranking them), which suffers from inevitable error propagation and limits the quality of the generated trailers. Beyond this paradigm, we propose a new self-paced and self-corrective masked prediction method called SSMP, which achieves state-of-the-art results in automatic trailer generation via bi-directional contextual modeling and progressive self-correction. In particular, SSMP trains a Transformer encoder that takes the movie shot sequences as prompts and generates corresponding trailer shot sequences accordingly. The model is trained via masked prediction, reconstructing each trailer shot sequence from its randomly masked counterpart. The mask ratio is self-paced, allowing the task difficulty to adapt to the model and thereby improving model performance. When generating a movie trailer, the model fills the shot positions with high confidence at each step and re-masks the remaining positions for the next prediction, forming a progressive self-correction mechanism that is analogous to how human editors work. Both quantitative results and user studies demonstrate the superiority of SSMP in comparison to existing automatic movie trailer generation methods. Demo is available at: https://github.com/Dixin-Lab/SSMP.

Paperid: 230, https://arxiv.org/pdf/2512.04425.pdf GitHub

Abstract:
Accurate and interpretable gait analysis plays a crucial role in the early detection of Parkinsons disease (PD),yet most existing approaches remain limited by single-modality inputs, low robustness, and a lack of clinical transparency. This paper presents an explainable multimodal framework that integrates RGB and Depth (RGB-D) data to recognize Parkinsonian gait patterns under realistic conditions. The proposed system employs dual YOLOv11-based encoders for modality-specific feature extraction, followed by a Multi-Scale Local-Global Extraction (MLGE) module and a Cross-Spatial Neck Fusion mechanism to enhance spatial-temporal representation. This design captures both fine-grained limb motion (e.g., reduced arm swing) and overall gait dynamics (e.g., short stride or turning difficulty), even in challenging scenarios such as low lighting or occlusion caused by clothing. To ensure interpretability, a frozen Large Language Model (LLM) is incorporated to translate fused visual embeddings and structured metadata into clinically meaningful textual explanations. Experimental evaluations on multimodal gait datasets demonstrate that the proposed RGB-D fusion framework achieves higher recognition accuracy, improved robustness to environmental variations, and clear visual-linguistic reasoning compared with single-input baselines. By combining multimodal feature learning with language-based interpretability, this study bridges the gap between visual recognition and clinical understanding, offering a novel vision-language paradigm for reliable and explainable Parkinsons disease gait analysis. Code:https://github.com/manaralnaasan/RGB-D_parkinson-LLM

Paperid: 231, https://arxiv.org/pdf/2512.04413.pdf GitHub

Abstract:
Knowledge distillation is an effective and hardware-friendly method, which plays a key role in lightweighting remote sensing object detection. However, existing distillation methods often encounter the issue of mixed features in remote sensing images (RSIs), and neglect the discrepancies caused by subtle feature variations, leading to entangled knowledge confusion. To address these challenges, we propose an architecture-agnostic distillation method named Dual-Stream Spectral Decoupling Distillation (DS2D2) for universal remote sensing object detection tasks. Specifically, DS2D2 integrates explicit and implicit distillation grounded in spectral decomposition. Firstly, the first-order wavelet transform is applied for spectral decomposition to preserve the critical spatial characteristics of RSIs. Leveraging this spatial preservation, a Density-Independent Scale Weight (DISW) is designed to address the challenges of dense and small object detection common in RSIs. Secondly, we show implicit knowledge hidden in subtle student-teacher feature discrepancies, which significantly influence predictions when activated by detection heads. This implicit knowledge is extracted via full-frequency and high-frequency amplifiers, which map feature differences to prediction deviations. Extensive experiments on DIOR and DOTA datasets validate the effectiveness of the proposed method. Specifically, on DIOR dataset, DS2D2 achieves improvements of 4.2% in AP50 for RetinaNet and 3.8% in AP50 for Faster R-CNN, outperforming existing distillation approaches. The source code will be available at https://github.com/PolarAid/DS2D2.

Paperid: 232, https://arxiv.org/pdf/2512.04390.pdf GitHub

Abstract:
Real-world video restoration is plagued by complex degradations from motion coupled with dynamically varying exposure - a key challenge largely overlooked by prior works and a common artifact of auto-exposure or low-light capture. We present FMA-Net++, a framework for joint video super-resolution and deblurring that explicitly models this coupled effect of motion and dynamically varying exposure. FMA-Net++ adopts a sequence-level architecture built from Hierarchical Refinement with Bidirectional Propagation blocks, enabling parallel, long-range temporal modeling. Within each block, an Exposure Time-aware Modulation layer conditions features on per-frame exposure, which in turn drives an exposure-aware Flow-Guided Dynamic Filtering module to infer motion- and exposure-aware degradation kernels. FMA-Net++ decouples degradation learning from restoration: the former predicts exposure- and motion-aware priors to guide the latter, improving both accuracy and efficiency. To evaluate under realistic capture conditions, we introduce REDS-ME (multi-exposure) and REDS-RE (random-exposure) benchmarks. Trained solely on synthetic data, FMA-Net++ achieves state-of-the-art accuracy and temporal consistency on our new benchmarks and GoPro, outperforming recent methods in both restoration quality and inference speed, and generalizes well to challenging real-world videos.

Paperid: 233, https://arxiv.org/pdf/2512.04356.pdf GitHub

Abstract:
Recent advancement in multimodal LLMs (MLLMs) has demonstrated their remarkable capability to generate descriptive captions for input videos. However, these models suffer from factual inaccuracies in the generated descriptions, causing severe hallucination issues. While prior works have explored alleviating hallucinations for static images, jointly mitigating visual object and temporal action hallucinations for dynamic videos remains a challenging and unsolved task. To tackle this challenge, we propose a Self-Augmented Contrastive Alignment (SANTA) framework for enabling object and action faithfulness by exempting the spurious correlations and enforcing the emphasis on visual facts. SANTA employs a hallucinative self-augmentation scheme to identify the potential hallucinations that lie in the MLLM and transform the original captions to the contrasted negatives. Furthermore, we develop a tracklet-phrase contrastive alignment to match the regional objects and relation-guided actions with their corresponding visual and temporal phrases. Extensive experiments demonstrate that SANTA outperforms existing methods in alleviating object and action hallucinations, yielding superior performance on the hallucination examination benchmarks.

Paperid: 234, https://arxiv.org/pdf/2512.04355.pdf GitHub

Abstract:
Modern GPU software stacks demand developers who can anticipate performance bottlenecks before ever launching a kernel; misjudging floating-point workloads upstream can derail tuning, scheduling, and even hardware procurement. Yet despite rapid progress in code generation, today's Large Language Models (LLMs) are rarely tested on this kind of forward-looking reasoning. We close that gap with gpuFLOPBench, a benchmark that asks models to "count without running" by predicting single and double-precision FLOP counts for 577 CUDA kernels drawn from HeCBench, annotated with ground-truth profiles and eight execution attributes that distinguish trivially analyzable code from kernels whose FLOPs depend on hidden compiler or runtime behavior. Evaluating current closed-source reasoning models shows clear but uneven progress: the newest LLMs achieve perfect classification on straightforward kernels but still incur multiple order-of-magnitude errors whenever implicit FLOPs arise from division, intrinsic math functions, or common subexpressions. These results surface a core limitation of existing code assistants -- the inability to internalize hardware-specific microcode effects -- and position gpuFLOPBench as a focused testbed for developing LLM tooling that can reason about performance with the same rigor as experienced GPU developers. Sources are available at our repository: https://github.com/Scientific-Computing-Lab/gpuFLOPBench

Paperid: 235, https://arxiv.org/pdf/2512.04305.pdf GitHub

Abstract:
While vision-language models like CLIP have been extensively studied, their calibration, crucial for reliable predictions, has received limited attention. Although a few prior works have examined CLIP calibration in offline settings, the impact of fine-tuning CLIP in a federated learning (FL) setup remains unexplored. In this work, we investigate how FL affects CLIP calibration and propose strategies to improve reliability in this distributed setting. We first analyze Textual Prompt Tuning approaches and show that they degrade calibration metrics when operating under FL. We also evaluate existing in-training calibration techniques across four global aggregation methods, finding that they provide limited improvements. Our results suggest that the key challenge lies not only in how we aggregate or calibrate, but in which components we choose to fine-tune. Motivated by this insight, we propose $\text{FL}^2\text{oRA}$, a straightforward LoRA-based approach that naturally improves calibration in FL, and we analyze the factors behind its effectiveness. Experiments on multiple benchmarks demonstrate that $\text{FL}^2\text{oRA}$ consistently produces well-calibrated models, reducing the need for explicit calibration procedures. Codes are available at https://github.com/mainaksingha01/FL2oRA.

Paperid: 236, https://arxiv.org/pdf/2512.04267.pdf GitHub

Abstract:
Lighting has a strong influence on visual appearance, yet understanding and representing lighting in images remains notoriously difficult. Various lighting representations exist, such as environment maps, irradiance, spherical harmonics, or text, but they are incompatible, which limits cross-modal transfer. We thus propose UniLight, a joint latent space as lighting representation, that unifies multiple modalities within a shared embedding. Modality-specific encoders for text, images, irradiance, and environment maps are trained contrastively to align their representations, with an auxiliary spherical-harmonics prediction task reinforcing directional understanding. Our multi-modal data pipeline enables large-scale training and evaluation across three tasks: lighting-based retrieval, environment-map generation, and lighting control in diffusion-based image synthesis. Experiments show that our representation captures consistent and transferable lighting features, enabling flexible manipulation across modalities.

Paperid: 237, https://arxiv.org/pdf/2512.04228.pdf GitHub

Abstract:
Large Language Models (LLMs) have transformed natural language processing and hold growing promise for advancing science, healthcare, and decision-making. Yet their training paradigms remain dominated by affirmation-based inference, akin to \textit{modus ponens}, where accepted premises yield predicted consequents. While effective for generative fluency, this one-directional approach leaves models vulnerable to logical fallacies, adversarial manipulation, and failures in causal reasoning. This paper makes two contributions. First, it demonstrates how existing LLMs from major platforms exhibit systematic weaknesses when reasoning in scientific domains with negation, counterexamples, or faulty premises \footnote{Code to recreate these experiments are at https://github.com/hannahdavidsoncollege-maker/ScientificReasoningForEnvironment-MedicineWithLLMs. Second, it introduces a dual-reasoning training framework that integrates affirmative generation with structured counterfactual denial. Grounded in formal logic, cognitive science, and adversarial training, this training paradigm formalizes a computational analogue of ``denying the antecedent'' as a mechanism for disconfirmation and robustness. By coupling generative synthesis with explicit negation-aware objectives, the framework enables models that not only affirm valid inferences but also reject invalid ones, yielding systems that are more resilient, interpretable, and aligned with human reasoning.

Paperid: 238, https://arxiv.org/pdf/2512.04213.pdf GitHub

Abstract:
This paper presents LAPA (Look Around and Pay Attention), a novel end-to-end transformer-based architecture for multi-camera point tracking that integrates appearance-based matching with geometric constraints. Traditional pipelines decouple detection, association, and tracking, leading to error propagation and temporal inconsistency in challenging scenarios. LAPA addresses these limitations by leveraging attention mechanisms to jointly reason across views and time, establishing soft correspondences through a cross-view attention mechanism enhanced with geometric priors. Instead of relying on classical triangulation, we construct 3D point representations via attention-weighted aggregation, inherently accommodating uncertainty and partial observations. Temporal consistency is further maintained through a transformer decoder that models long-range dependencies, preserving identities through extended occlusions. Extensive experiments on challenging datasets, including our newly created multi-camera (MC) versions of TAPVid-3D panoptic and PointOdyssey, demonstrate that our unified approach significantly outperforms existing methods, achieving 37.5% APD on TAPVid-3D-MC and 90.3% APD on PointOdyssey-MC, particularly excelling in scenarios with complex motions and occlusions. Code is available at https://github.com/ostadabbas/Look-Around-and-Pay-Attention-LAPA-

Paperid: 239, https://arxiv.org/pdf/2512.04194.pdf GitHub

Abstract:
We propose a simple safety filter design for stochastic discrete-time systems based on piecewise affine probabilistic control barrier functions, providing an appealing balance between modeling flexibility and computational complexity. Exact evaluation of the safety filter consists of solving a mixed-integer quadratic program (MIQP) if the dynamics are control-affine (or a mixed-integer nonlinear program in general). We propose a heuristic search method that replaces this by a small number of small-scale quadratic programs (QPs), or nonlinear programs (NLPs) respectively. The proposed approach provides a flexible framework in which arbitrary (data-driven) quantile estimators can be used to bound the probability of safety violations. Through extensive numerical experiments, we demonstrate improvements in conservatism and computation time with respect to existing methods, and we illustrate the flexibility of the method for modeling complex safety sets. Supplementary material can be found at https://mathijssch.github.io/ecc26-supplementary/.

Paperid: 240, https://arxiv.org/pdf/2512.04169.pdf GitHub

Abstract:
Lattice surgery with two-dimensional quantum error correcting codes is among the leading schemes for fault-tolerant quantum computation, motivated by superconducting hardware architectures. In conventional lattice surgery compilation schemes, logical circuits are compiled following a place-and-route paradigm, where logical qubits remain statically fixed in space throughout the computation. In this work, we introduce a paradigm shift by exploiting movable logical qubits via teleportation during the logical lattice surgery CNOT gate. Focusing on lattice surgery with the color code, we propose a proof-of-concept compilation scheme that leverages this capability. Numerical simulations show that the proposed approach can substantially reduce the routed circuit depth compared to standard place-and-route compilation techniques. Our results demonstrate that optimizations based on movable logical qubits are not limited to architectures with physically movable qubits, such as neutral atoms or trapped ions - they are also readily applicable to superconducting quantum hardware. An open-source implementation of our method is available on GitHub https://github.com/munich-quantum-toolkit/qecc.

Paperid: 241, https://arxiv.org/pdf/2512.04125.pdf GitHub

Abstract:
Large language models (LLMs) have demonstrated several emergent behaviors with scale, including reasoning and fluency in long-form text generation. However, they continue to struggle with tasks requiring precise spatial and positional reasoning. ASCII art, a symbolic medium where characters encode structure and form, provides a unique probe of this limitation. We introduce ASCIIBench, a novel benchmark for evaluating both the generation and classification of ASCII-text images. ASCIIBench consists of a filtered dataset of 5,315 class-labeled ASCII images and is, to our knowledge, the first publicly available benchmark of its kind. Alongside the dataset, we release weights for a fine-tuned CLIP model adapted to capture ASCII structure, enabling the evaluation of LLM-generated ASCII art. Our analysis shows that cosine similarity over CLIP embeddings fails to separate most ASCII categories, yielding chance-level performance even for low-variance classes. In contrast, classes with high internal mean similarity exhibit clear discriminability, revealing that the bottleneck lies in representation rather than generational variance. These findings position ASCII art as a stress test for multimodal representations and motivate the development of new embedding methods or evaluation metrics tailored to symbolic visual modalities. All resources are available at https://github.com/ASCIIBench/ASCIIBench.

Paperid: 242, https://arxiv.org/pdf/2512.04120.pdf GitHub

Abstract:
The emergence of open data portals necessitates more attention to protecting sensitive data before datasets get published and exchanged. While an abundance of methods for suppressing sensitive data exist, the conceptualization of sensitive data and methods to detect it, focus particularly on personal data that, if disclosed, may be harmful or violate privacy. We observe the need for refining and broadening our definitions of sensitive data, and argue that the sensitivity of data depends on its context. Based on this definition, we introduce two mechanisms for contextual sensitive data detection that consider the broader context of a dataset at hand. First, we introduce type contextualization, which first detects the semantic type of particular data values, then considers the overall context of the data values within the dataset or document. Second, we introduce domain contextualization which determines sensitivity of a given dataset in the broader context based on the retrieval of relevant rules from documents that specify data sensitivity (e.g., data topic and geographic origin). Experiments with these mechanisms, assisted by large language models (LLMs), confirm that: 1) type-contextualization significantly reduces the number of false positives for type-based sensitive data detection and reaches a recall of 94% compared to 63% with commercial tools, and 2) domain-contextualization leveraging sensitivity rule retrieval is effective for context-grounded sensitive data detection in non-standard data domains such as humanitarian datasets. Evaluation with humanitarian data experts also reveals that context-grounded LLM explanations provide useful guidance in manual data auditing processes, improving consistency. We open-source mechanisms and annotated datasets for contextual sensitive data detection at https://github.com/trl-lab/sensitive-data-detection.

Paperid: 243, https://arxiv.org/pdf/2512.04084.pdf GitHub

Abstract:
Normalizing Flows (NFs) learn invertible mappings between the data and a Gaussian distribution. Prior works usually suffer from two limitations. First, they add random noise to training samples or VAE latents as data augmentation, introducing complex pipelines including extra noising and denoising steps. Second, they use a pretrained and frozen VAE encoder, resulting in suboptimal reconstruction and generation quality. In this paper, we find that the two issues can be solved in a very simple way: just fixing the variance (which would otherwise be predicted by the VAE encoder) to a constant (e.g., 0.5). On the one hand, this method allows the encoder to output a broader distribution of tokens and the decoder to learn to reconstruct clean images from the augmented token distribution, avoiding additional noise or denoising design. On the other hand, fixed variance simplifies the VAE evidence lower bound, making it stable to train an NF with a VAE jointly. On the ImageNet $256 \times 256$ generation task, our model SimFlow obtains a gFID score of 2.15, outperforming the state-of-the-art method STARFlow (gFID 2.40). Moreover, SimFlow can be seamlessly integrated with the end-to-end representation alignment (REPA-E) method and achieves an improved gFID of 1.91, setting a new state of the art among NFs.

Paperid: 244, https://arxiv.org/pdf/2512.04069.pdf GitHub

Abstract:
Vision Language Models (VLMs) demonstrate strong qualitative visual understanding, but struggle with metrically precise spatial reasoning required for embodied applications. The agentic paradigm promises that VLMs can use a wide variety of tools that could augment these capabilities, such as depth estimators, segmentation models, and pose estimators. Yet it remains an open challenge how to realize this vision without solely relying on handcrafted prompting strategies or enforcing fixed, predefined tool pipelines that limit VLMs' ability to discover optimal tool-use patterns. Reinforcement Learning could overcome this gap, but has so far been limited to reasoning with a single visual tool due to the large search space in multi-tool reasoning. We introduce Double Interactive Reinforcement Learning (DIRL), a two-phase training framework where VLMs learn to coordinate multiple tools through interactive exploration and feedback. In the teaching phase, we combine demonstrations from a single tool specialist trained via interactive RL with traces from a frontier model using all tools. In the exploration phase, the model further refines multi-tool coordination through continued RL. Our model, SpaceTools, with tool-augmented spatial reasoning ability, achieves state-of-the-art performance on spatial understanding benchmarks (RoboSpatial-Home, BLINK, BOP-ASK) and demonstrates reliable real-world manipulation using a 7-DOF robot as a tool. DIRL provides substantial improvements over the vanilla SFT (+12% on RoboSpatial) and RL (+16% on RoboSpatial) baselines. Project page: https://spacetools.github.io/.

Paperid: 245, https://arxiv.org/pdf/2512.04022.pdf GitHub

Abstract:
This study investigates the non-linear determinants of pedestrian injury severity using administrative data from Great Britain's 2023 STATS19 dataset. To address inherent data-quality challenges, including missing information and substantial class imbalance, we employ a rigorous preprocessing pipeline utilizing mode imputation and Synthetic Minority Over-sampling (SMOTE). We utilize non-parametric ensemble methods (Random Forest and XGBoost) to capture complex interactions and heterogeneity often missed by linear models, while Shapley Additive Explanations are employed to ensure interpretability and isolate marginal feature effects. Our analysis reveals that vehicle count, speed limits, lighting, and road surface conditions are the primary predictors of severity, with police attendance and junction characteristics further distinguishing severe collisions. Spatially, while pedestrian risk is concentrated in dense urban Local Authority Districts (LADs), we identify that certain rural LADs experience disproportionately severe outcomes conditional on a collision occurring. These findings underscore the value of combining spatial analysis with interpretable machine learning to guide geographically targeted speed management, infrastructure investment, and enforcement strategies.

Paperid: 246, https://arxiv.org/pdf/2512.04021.pdf GitHub

Abstract:
Reconstructing and understanding 3D scenes from unposed sparse views in a feed-forward manner remains as a challenging task in 3D computer vision. Recent approaches use per-pixel 3D Gaussian Splatting for reconstruction, followed by a 2D-to-3D feature lifting stage for scene understanding. However, they generate excessive redundant Gaussians, causing high memory overhead and sub-optimal multi-view feature aggregation, leading to degraded novel view synthesis and scene understanding performance. We propose C3G, a novel feed-forward framework that estimates compact 3D Gaussians only at essential spatial locations, minimizing redundancy while enabling effective feature lifting. We introduce learnable tokens that aggregate multi-view features through self-attention to guide Gaussian generation, ensuring each Gaussian integrates relevant visual features across views. We then exploit the learned attention patterns for Gaussian decoding to efficiently lift features. Extensive experiments on pose-free novel view synthesis, 3D open-vocabulary segmentation, and view-invariant feature aggregation demonstrate our approach's effectiveness. Results show that a compact yet geometrically meaningful representation is sufficient for high-quality scene reconstruction and understanding, achieving superior memory efficiency and feature fidelity compared to existing methods.

Paperid: 247, https://arxiv.org/pdf/2512.04015.pdf GitHub

Abstract:
Modeling group actions on latent representations enables controllable transformations of high-dimensional image data. Prior works applying group-theoretic priors or modeling transformations typically operate in the high-dimensional data space, where group actions apply uniformly across the entire input, making it difficult to disentangle the subspace that varies under transformations. While latent-space methods offer greater flexibility, they still require manual partitioning of latent variables into equivariant and invariant subspaces, limiting the ability to robustly learn and operate group actions within the representation space. To address this, we introduce a novel end-to-end framework that for the first time learns group actions on latent image manifolds, automatically discovering transformation-relevant structures without manual intervention. Our method uses learnable binary masks with straight-through estimation to dynamically partition latent representations into transformation-sensitive and invariant components. We formulate this within a unified optimization framework that jointly learns latent disentanglement and group transformation mappings. The framework can be seamlessly integrated with any standard encoder-decoder architecture. We validate our approach on five 2D/3D image datasets, demonstrating its ability to automatically learn disentangled latent factors for group actions in diverse data, while downstream classification tasks confirm the effectiveness of the learned representations. Our code is publicly available at https://github.com/farhanaswarnali/Learning-Group-Actions-In-Disentangled-Latent-Image-Representations .

Paperid: 248, https://arxiv.org/pdf/2512.04012.pdf GitHub

Abstract:
Reliable 3D reconstruction from in-the-wild image collections is often hindered by "noisy" images-irrelevant inputs with little or no view overlap with others. While traditional Structure-from-Motion pipelines handle such cases through geometric verification and outlier rejection, feed-forward 3D reconstruction models lack these explicit mechanisms, leading to degraded performance under in-the-wild conditions. In this paper, we discover that the existing feed-forward reconstruction model, e.g., VGGT, despite lacking explicit outlier-rejection mechanisms or noise-aware training, can inherently distinguish distractor images. Through an in-depth analysis under varying proportions of synthetic distractors, we identify a specific layer that naturally exhibits outlier-suppressing behavior. Further probing reveals that this layer encodes discriminative internal representations that enable an effective noise-filtering capability, which we simply leverage to perform outlier-view rejection in feed-forward 3D reconstruction without any additional fine-tuning or supervision. Extensive experiments on both controlled and in-the-wild datasets demonstrate that this implicit filtering mechanism is consistent and generalizes well across diverse scenarios.

Paperid: 249, https://arxiv.org/pdf/2512.03996.pdf GitHub GitHub

Abstract:
Test-time scaling (TTS) aims to achieve better results by increasing random sampling and evaluating samples based on rules and metrics. However, in text-to-image(T2I) diffusion models, most related works focus on search strategies and reward models, yet the impact of the stochastic characteristic of noise in T2I diffusion models on the method's performance remains unexplored. In this work, we analyze the effects of randomness in T2I diffusion models and explore a new format of randomness for TTS: text embedding perturbation, which couples with existing randomness like SDE-injected noise to enhance generative diversity and quality. We start with a frequency-domain analysis of these formats of randomness and their impact on generation, and find that these two randomness exhibit complementary behavior in the frequency domain: spatial noise favors low-frequency components (early steps), while text embedding perturbation enhances high-frequency details (later steps), thereby compensating for the potential limitations of spatial noise randomness in high-frequency manipulation. Concurrently, text embedding demonstrates varying levels of tolerance to perturbation across different dimensions of the generation process. Specifically, our method consists of two key designs: (1) Introducing step-based text embedding perturbation, combining frequency-guided noise schedules with spatial noise perturbation. (2) Adapting the perturbation intensity selectively based on their frequency-specific contributions to generation and tolerance to perturbation. Our approach can be seamlessly integrated into existing TTS methods and demonstrates significant improvements on multiple benchmarks with almost no additional computation. Code is available at \href{https://github.com/xuhang07/TEP-Diffusion}{https://github.com/xuhang07/TEP-Diffusion}.

Paperid: 250, https://arxiv.org/pdf/2512.03981.pdf GitHub GitHub

Abstract:
Drag-based image editing using generative models provides intuitive control over image structures. However, existing methods rely heavily on manually provided masks and textual prompts to preserve semantic fidelity and motion precision. Removing these constraints creates a fundamental trade-off: visual artifacts without masks and poor spatial control without prompts. To address these limitations, we propose DirectDrag, a novel mask- and prompt-free editing framework. DirectDrag enables precise and efficient manipulation with minimal user input while maintaining high image fidelity and accurate point alignment. DirectDrag introduces two key innovations. First, we design an Auto Soft Mask Generation module that intelligently infers editable regions from point displacement, automatically localizing deformation along movement paths while preserving contextual integrity through the generative model's inherent capacity. Second, we develop a Readout-Guided Feature Alignment mechanism that leverages intermediate diffusion activations to maintain structural consistency during point-based edits, substantially improving visual fidelity. Despite operating without manual mask or prompt, DirectDrag achieves superior image quality compared to existing methods while maintaining competitive drag accuracy. Extensive experiments on DragBench and real-world scenarios demonstrate the effectiveness and practicality of DirectDrag for high-quality, interactive image manipulation. Project Page: https://frakw.github.io/DirectDrag/. Code is available at: https://github.com/frakw/DirectDrag.

Paperid: 251, https://arxiv.org/pdf/2512.03979.pdf GitHub

Abstract:
Diffusion models show promise for dynamic scene deblurring; however, existing studies often fail to leverage the intrinsic nature of the blurring process within diffusion models, limiting their full potential. To address it, we present a Blur Diffusion Model (BlurDM), which seamlessly integrates the blur formation process into diffusion for image deblurring. Observing that motion blur stems from continuous exposure, BlurDM implicitly models the blur formation process through a dual-diffusion forward scheme, diffusing both noise and blur onto a sharp image. During the reverse generation process, we derive a dual denoising and deblurring formulation, enabling BlurDM to recover the sharp image by simultaneously denoising and deblurring, given pure Gaussian noise conditioned on the blurred image as input. Additionally, to efficiently integrate BlurDM into deblurring networks, we perform BlurDM in the latent space, forming a flexible prior generation network for deblurring. Extensive experiments demonstrate that BlurDM significantly and consistently enhances existing deblurring methods on four benchmark datasets. The source code is available at https://github.com/Jin-Ting-He/BlurDM.

Paperid: 252, https://arxiv.org/pdf/2512.03973.pdf GitHub

Abstract:
Offline reinforcement learning often relies on behavior regularization that enforces policies to remain close to the dataset distribution. However, such approaches fail to distinguish between high-value and low-value actions in their regularization components. We introduce Guided Flow Policy (GFP), which couples a multi-step flow-matching policy with a distilled one-step actor. The actor directs the flow policy through weighted behavior cloning to focus on cloning high-value actions from the dataset rather than indiscriminately imitating all state-action pairs. In turn, the flow policy constrains the actor to remain aligned with the dataset's best transitions while maximizing the critic. This mutual guidance enables GFP to achieve state-of-the-art performance across 144 state and pixel-based tasks from the OGBench, Minari, and D4RL benchmarks, with substantial gains on suboptimal datasets and challenging tasks. Webpage: https://simple-robotics.github.io/publications/guided-flow-policy/

Paperid: 253, https://arxiv.org/pdf/2512.03964.pdf GitHub

Abstract:
Tuning-free face personalization methods have developed along two distinct paradigms: text embedding approaches that map facial features into the text embedding space, and adapter-based methods that inject features through auxiliary cross-attention layers. While both paradigms have shown promise, existing methods struggle to simultaneously achieve high identity fidelity and flexible text controllability. We introduce UniID, a unified tuning-free framework that synergistically integrates both paradigms. Our key insight is that when merging these approaches, they should mutually reinforce only identity-relevant information while preserving the original diffusion prior for non-identity attributes. We realize this through a principled training-inference strategy: during training, we employ an identity-focused learning scheme that guides both branches to capture identity features exclusively; at inference, we introduce a normalized rescaling mechanism that recovers the text controllability of the base diffusion model while enabling complementary identity signals to enhance each other. This principled design enables UniID to achieve high-fidelity face personalization with flexible text controllability. Extensive experiments against six state-of-the-art methods demonstrate that UniID achieves superior performance in both identity preservation and text controllability. Code will be available at https://github.com/lyuPang/UniID

Paperid: 254, https://arxiv.org/pdf/2512.03958.pdf GitHub

Abstract:
Agricultural robots are serving as powerful assistants across a wide range of agricultural tasks, nevertheless, still heavily relying on manual operations or railway systems for movement. The AgriVLN method and the A2A benchmark pioneeringly extend Vision-and-Language Navigation (VLN) to the agricultural domain, enabling a robot to navigate to a target position following a natural language instruction. Unlike human binocular vision, most agricultural robots are only given a single camera for monocular vision, which results in limited spatial perception. To bridge this gap, we present the method of Agricultural Vision-and-Language Navigation with Monocular Depth Estimation (MDE-AgriVLN), in which we propose the MDE module generating depth features from RGB images, to assist the decision-maker on reasoning. When evaluated on the A2A benchmark, our MDE-AgriVLN method successfully increases Success Rate from 0.23 to 0.32 and decreases Navigation Error from 4.43m to 4.08m, demonstrating the state-of-the-art performance in the agricultural VLN domain. Code: https://github.com/AlexTraveling/MDE-AgriVLN.

Paperid: 255, https://arxiv.org/pdf/2512.03932.pdf GitHub

Abstract:
Deep learning-based image restoration has achieved significant success. However, when addressing real-world degradations, model performance is limited by the quality of ground-truth images in datasets due to practical constraints in data acquisition. To address this limitation, we propose a novel framework that enhances existing ground truth images to provide higher-quality supervision for real-world restoration. Our framework generates perceptually enhanced ground truth images using super-resolution by incorporating adaptive frequency masks, which are learned by a conditional frequency mask generator. These masks guide the optimal fusion of frequency components from the original ground truth and its super-resolved variants, yielding enhanced ground truth images. This frequency-domain mixup preserves the semantic consistency of the original content while selectively enriching perceptual details, preventing hallucinated artifacts that could compromise fidelity. The enhanced ground truth images are used to train a lightweight output refinement network that can be seamlessly integrated with existing restoration models. Extensive experiments demonstrate that our approach consistently improves the quality of restored images. We further validate the effectiveness of both supervision enhancement and output refinement through user studies. Code is available at https://github.com/dhryougit/Beyond-the-Ground-Truth.

Paperid: 256, https://arxiv.org/pdf/2512.03918.pdf GitHub

Abstract:
We propose UniMo, an innovative autoregressive model for joint modeling of 2D human videos and 3D human motions within a unified framework, enabling simultaneous generation and understanding of these two modalities for the first time. Current methods predominantly focus on generating one modality given another as the condition or integrating either of them with other modalities such as text and audio. Unifying 2D videos and 3D motions for simultaneous optimization and generation remains largely unexplored, presenting significant challenges due to their substantial structural and distributional differences. Inspired by the LLM's ability to unify different modalities, our method models videos and 3D motions as a unified tokens sequence, utilizing separate embedding layers to mitigate distribution gaps. Additionally, we devise a sequence modeling strategy that integrates two distinct tasks within a single framework, proving the effectiveness of unified modeling. Moreover, to efficiently align with visual tokens and preserve 3D spatial information, we design a novel 3D motion tokenizer with a temporal expansion strategy, using a single VQ-VAE to produce quantized motion tokens. It features multiple expert decoders that handle body shapes, translation, global orientation, and body poses for reliable 3D motion reconstruction. Extensive experiments demonstrate that our method simultaneously generates corresponding videos and motions while performing accurate motion capture. This work taps into the capacity of LLMs to fuse diverse data types, paving the way for integrating human-centric information into existing models and potentially enabling multimodal, controllable joint modeling of humans, objects, and scenes.

Paperid: 257, https://arxiv.org/pdf/2512.03905.pdf GitHub GitHub

Abstract:
The remarkable success in text-to-image diffusion models has motivated extensive investigation of their potential for video applications. Zero-shot techniques aim to adapt image diffusion models for videos without requiring further model training. Recent methods largely emphasize integrating inter-frame correspondence into attention mechanisms. However, the soft constraint applied to identify the valid features to attend is insufficient, which could lead to temporal inconsistency. In this paper, we present FRESCO, which integrates intra-frame correspondence with inter-frame correspondence to formulate a more robust spatial-temporal constraint. This enhancement ensures a consistent transformation of semantically similar content between frames. Our method goes beyond attention guidance to explicitly optimize features, achieving high spatial-temporal consistency with the input video, significantly enhancing the visual coherence of manipulated videos. We verify FRESCO adaptations on two zero-shot tasks of video-to-video translation and text-guided video editing. Comprehensive experiments demonstrate the effectiveness of our framework in generating high-quality, coherent videos, highlighting a significant advance over current zero-shot methods.

Paperid: 258, https://arxiv.org/pdf/2512.03887.pdf GitHub GitHub

Abstract:
The advancement in Large Language Models has driven the creation of complex agentic systems, such as Deep Research Agents (DRAs), to overcome the limitations of static Retrieval Augmented Generation (RAG) pipelines in handling complex, multi-turn research tasks. This paper introduces the Static Deep Research Agent (Static-DRA), a novel solution built upon a configurable and hierarchical Tree-based static workflow. The core contribution is the integration of two user-tunable parameters, Depth and Breadth, which provide granular control over the research intensity. This design allows end-users to consciously balance the desired quality and comprehensiveness of the research report against the associated computational cost of Large Language Model (LLM) interactions. The agent's architecture, comprising Supervisor, Independent, and Worker agents, facilitates effective multi-hop information retrieval and parallel sub-topic investigation. We evaluate the Static-DRA against the established DeepResearch Bench using the RACE (Reference-based Adaptive Criteria-driven Evaluation) framework. Configured with a depth of 2 and a breadth of 5, and powered by the gemini-2.5-pro model, the agent achieved an overall score of 34.72. Our experiments validate that increasing the configured Depth and Breadth parameters results in a more in-depth research process and a correspondingly higher evaluation score. The Static-DRA offers a pragmatic and resource-aware solution, empowering users with transparent control over the deep research process. The entire source code, outputs and benchmark results are open-sourced at https://github.com/SauravP97/Static-Deep-Research/

Paperid: 259, https://arxiv.org/pdf/2512.03844.pdf GitHub

Abstract:
Prevailing Dataset Distillation (DD) methods leveraging generative models confront two fundamental limitations. First, despite pioneering the use of diffusion models in DD and delivering impressive performance, the vast majority of approaches paradoxically require a diffusion model pre-trained on the full target dataset, undermining the very purpose of DD and incurring prohibitive training costs. Second, although some methods turn to general text-to-image models without relying on such target-specific training, they suffer from a significant distributional mismatch, as the web-scale priors encapsulated in these foundation models fail to faithfully capture the target-specific semantics, leading to suboptimal performance. To tackle these challenges, we propose Core Distribution Alignment (CoDA), a framework that enables effective DD using only an off-the-shelf text-to-image model. Our key idea is to first identify the "intrinsic core distribution" of the target dataset using a robust density-based discovery mechanism. We then steer the generative process to align the generated samples with this core distribution. By doing so, CoDA effectively bridges the gap between general-purpose generative priors and target semantics, yielding highly representative distilled datasets. Extensive experiments suggest that, without relying on a generative model specifically trained on the target dataset, CoDA achieves performance on par with or even superior to previous methods with such reliance across all benchmarks, including ImageNet-1K and its subsets. Notably, it establishes a new state-of-the-art accuracy of 60.4% at the 50-images-per-class (IPC) setup on ImageNet-1K. Our code is available on the project webpage: https://github.com/zzzlt422/CoDA

Paperid: 260, https://arxiv.org/pdf/2512.03837.pdf GitHub

Abstract:
Human action recognition (HAR) in videos has garnered widespread attention due to the rich information in RGB videos. Nevertheless, existing methods for extracting deep features from RGB videos face challenges such as information redundancy, susceptibility to noise and high storage costs. To address these issues and fully harness the useful information in videos, we propose a novel heatmap pooling network (HP-Net) for action recognition from videos, which extracts information-rich, robust and concise pooled features of the human body in videos through a feedback pooling module. The extracted pooled features demonstrate obvious performance advantages over the previously obtained pose data and heatmap features from videos. In addition, we design a spatial-motion co-learning module and a text refinement modulation module to integrate the extracted pooled features with other multimodal data, enabling more robust action recognition. Extensive experiments on several benchmarks namely NTU RGB+D 60, NTU RGB+D 120, Toyota-Smarthome and UAV-Human consistently verify the effectiveness of our HP-Net, which outperforms the existing human action recognition methods. Our code is publicly available at: https://github.com/liujf69/HPNet-Action.

Paperid: 261, https://arxiv.org/pdf/2512.03759.pdf GitHub

Abstract:
Reinforcement Learning (RL) has proven highly effective for autoregressive language models, but adapting these methods to diffusion large language models (dLLMs) presents fundamental challenges. The core difficulty lies in likelihood approximation: while autoregressive models naturally provide token-level conditional probabilities essential for token-level RL objectives (e.g., GRPO), dLLMs generate sequences through iterative non-autoregressive denoising steps that lack this factorization. To address this fundamental mismatch, we propose ELBO-based Sequence-level Policy Optimization (ESPO), a principled RL framework that treats entire sequence generation as a single action and uses the ELBO as a tractable sequence-level likelihood proxy. Our method incorporates per-token normalization of importance ratios and robust KL-divergence estimation to ensure stable large-scale training. Extensive experiments on mathematical reasoning, coding, and planning tasks demonstrate that ESPO significantly outperforms token-level baselines, achieving dramatic improvements of 20-40 points on the Countdown task, while maintaining consistent gains on math and coding benchmarks. Our approach establishes sequence-level optimization as a principled and empirically effective paradigm for RL in dLLMs. Our code is available at https://github.com/ML-GSAI/ESPO.

Paperid: 262, https://arxiv.org/pdf/2512.03756.pdf GitHub

Abstract:
Combining motion prediction and motion planning offers a promising framework for enhancing interactions between automated vehicles and other traffic participants. However, this introduces challenges in conditioning predictions on navigation goals and ensuring stable, kinematically feasible trajectories. Addressing the former challenge, this paper investigates the extension of attention-based motion prediction models with navigation information. By integrating the ego vehicle's intended route and goal pose into the model architecture, we bridge the gap between multi-agent motion prediction and goal-based motion planning. We propose and evaluate several architectural navigation integration strategies to our model on the nuPlan dataset. Our results demonstrate the potential of prediction-driven motion planning, highlighting how navigation information can enhance both prediction and planning tasks. Our implementation is at: https://github.com/KIT-MRT/future-motion.

Paperid: 263, https://arxiv.org/pdf/2512.03746.pdf GitHub

Abstract:
Multimodal large language models (MLLMs) that think with images can interactively use tools to reason about visual inputs, but current approaches often rely on a narrow set of tools with limited real-world necessity and scalability. In this work, we first reveal a critical and previously overlooked weakness: even state-of-the-art MLLMs are surprisingly brittle, showing significant performance degradation on images with simple orientation changes or natural corruptions, underscoring the need for more robust tool-based reasoning. To address this, we propose CodeVision, a flexible and scalable code-as-tool framework where the model generates code as a universal interface to invoke any image operation, moving beyond fixed tool registries. We train our model using a two-stage methodology, beginning with Supervised Fine-Tuning (SFT) on a high-quality dataset curated for complex, multi-turn tool composition and error recovery, followed by Reinforcement Learning (RL) with a novel and dense process reward function to encourage strategic and efficient tool use. To facilitate this research, we construct new SFT and RL datasets and introduce a challenging new benchmark suite designed to rigorously evaluate robustness to orientation changes and multi-tool reasoning. Experiments on Qwen2.5-VL and Qwen3-VL series show that our approach significantly improves model performance and fosters emergent capabilities such as flexible tool composition, efficient chained execution, and robust error recovery from runtime feedback. Code is available at https://github.com/ByteDance-BandAI/CodeVision.

Paperid: 264, https://arxiv.org/pdf/2512.03720.pdf GitHub

Abstract:
Large Language Models (LLMs) have emerged as powerful tools for diverse applications. However, their uniform token processing paradigm introduces critical vulnerabilities in instruction handling, particularly when exposed to adversarial scenarios. In this work, we identify and propose a novel class of vulnerabilities, termed Tool-Completion Attack (TCA), which exploits function-calling mechanisms to subvert model behavior. To evaluate LLM robustness against such threats, we introduce the Tool-Completion benchmark, a comprehensive security assessment framework, which reveals that even state-of-the-art models remain susceptible to TCA, with surprisingly high attack success rates. To address these vulnerabilities, we introduce Context-Aware Hierarchical Learning (CAHL), a sophisticated mechanism that dynamically balances semantic comprehension with role-specific instruction constraints. CAHL leverages the contextual correlations between different instruction segments to establish a robust, context-aware instruction hierarchy. Extensive experiments demonstrate that CAHL significantly enhances LLM robustness against both conventional attacks and the proposed TCA, exhibiting strong generalization capabilities in zero-shot evaluations while still preserving model performance on generic tasks. Our code is available at https://github.com/S2AILab/CAHL.

Paperid: 265, https://arxiv.org/pdf/2512.03704.pdf GitHub

Abstract:
Long-context dialogue systems suffer from State Inertia, where static constraints prevent models from resolving conflicts between evolving user intents and established historical context. To address this, we propose DZ-TDPO, a non-destructive alignment framework that synergizes conflict-aware dynamic KL constraints with a calibrated temporal attention bias. Experiments on the Multi-Session Chat (MSC) dataset demonstrate that DZ-TDPO achieves state-of-the-art win rates (55.4% on Phi-3.5) while maintaining robust zero-shot generalization. Our scaling analysis reveals a "Capacity-Stability Trade-off": while smaller models incur an "alignment tax" (perplexity surge) to overcome historical inertia, the larger Qwen2.5-7B model achieves 50.8% win rate with negligible perplexity overhead. This confirms that TAI can be alleviated via precise attention regulation rather than destructive weight updates, preserving general capabilities (MMLU) across model scales. Code and data are available: https://github.com/lyj20071013/DZ-TDPO

Paperid: 266, https://arxiv.org/pdf/2512.03672.pdf GitHub

Abstract:
Hydro-Science and Engineering (Hydro-SE) is a critical and irreplaceable domain that secures human water supply, generates clean hydropower energy, and mitigates flood and drought disasters. Featuring multiple engineering objectives, Hydro-SE is an inherently interdisciplinary domain that integrates scientific knowledge with engineering expertise. This integration necessitates extensive expert collaboration in decision-making, which poses difficulties for intelligence. With the rapid advancement of large language models (LLMs), their potential application in the Hydro-SE domain is being increasingly explored. However, the knowledge and application abilities of LLMs in Hydro-SE have not been sufficiently evaluated. To address this issue, we propose the Hydro-SE LLM evaluation benchmark (Hydro-SE Bench), which contains 4,000 multiple-choice questions. Hydro-SE Bench covers nine subfields and enables evaluation of LLMs in aspects of basic conceptual knowledge, engineering application ability, and reasoning and calculation ability. The evaluation results on Hydro-SE Bench show that the accuracy values vary among 0.74 to 0.80 for commercial LLMs, and among 0.41 to 0.68 for small-parameter LLMs. While LLMs perform well in subfields closely related to natural and physical sciences, they struggle with domain-specific knowledge such as industry standards and hydraulic structures. Model scaling mainly improves reasoning and calculation abilities, but there is still great potential for LLMs to better handle problems in practical engineering application. This study highlights the strengths and weaknesses of LLMs for Hydro-SE tasks, providing model developers with clear training targets and Hydro-SE researchers with practical guidance for applying LLMs.

Paperid: 267, https://arxiv.org/pdf/2512.03667.pdf GitHub

Abstract:
In this study, we present Colon-X, an open initiative aimed at advancing multimodal intelligence in colonoscopy. We begin by constructing ColonVQA, the most comprehensive multimodal dataset ever built for colonoscopy, featuring over 1.1M+ visual question answering entries across 76 clinical findings and 18 multimodal tasks. Beyond serving as a community-wide data foundation, we further investigate a critical yet underexplored transition in colonoscopy - evolving from multimodal understanding to clinical reasoning: (a) To capture the current landscape of multimodal understanding behaviors, we systematically assess the generalizability of 22 multimodal large language models and examine their reliability under human-induced perturbations. The results reveal that clinical outputs from leading MLLMs remain far from robust and trustworthy. (b) To narrow this gap, we further explore reasoning-centric intelligence tailored for colonoscopy. Specifically, we curate ColonReason, a clinically grounded reasoning dataset annotated through a multi-expert debating pipeline, and develop ColonR1, the first R1-styled model incorporating task-adaptive rewarding and gradient-stable optimization techniques. Under data-scarce conditions, our ColonR1 achieves 56.61% overall accuracy, outperforming supervised fine-tuning by 25.22%, and sets a new reasoning-enabled baseline for multimodal colonoscopy analysis. All data and model resources are publicly available at https://github.com/ai4colonoscopy/Colon-X.

Paperid: 268, https://arxiv.org/pdf/2512.03666.pdf GitHub GitHub

Abstract:
A core capability towards general embodied intelligence lies in localizing task-relevant objects from an egocentric perspective, formulated as Spatio-Temporal Video Grounding (STVG). Despite recent progress, existing STVG studies remain largely confined to object-centric and descriptive instructions, neglecting the task-oriented reasoning that is crucial for embodied agents to accomplish goal-directed interactions. To bridge this gap, we introduce \textbf{ToG-Bench}, the first task-oriented spatio-temporal video grounding benchmark for egocentric videos. ToG-Bench is characterized by three key features: (1) \textbf{Task-oriented Grounding}, which requires identifying and localizing objects based on intended tasks rather than straightforward descriptions; (2) \textbf{Explicit-Implicit Dual Grounding}, where target objects can be either explicitly mentioned or implicitly inferred by contextual reasoning; (3) \textbf{One-to-Many Grounding}, where a single instruction may correspond to multiple objects involved in task execution. Built upon videos sourced from ScanNet, ToG-Bench comprises 100 annotated clips with 2,704 task-oriented grounding instructions, constructed via a semi-automated pipeline that combines foundation model annotation and human refinement. In addition, we introduce a set of task-level evaluation metrics tailored for multi-object and explicit-implicit object grounding, and systematically benchmark seven state-of-the-art MLLMs. Extensive experiments reveal the intrinsic challenges of task-oriented STVG and substantial performance gaps across explicit-implicit and multi-object grounding, highlighting the difficulty of bridging perception and interaction in embodied scenarios. Data and code will be released at: \href{https://github.com/qaxuDev/ToG-Bench}{https://github.com/qaxuDev/ToG-Bench}..

Paperid: 269, https://arxiv.org/pdf/2512.03643.pdf GitHub

Abstract:
DeepSeek-OCR demonstrates that rendered text can be reconstructed with high fidelity from a small number of vision tokens. This finding has sparked excitement about vision-based context compression for language models. But the evaluation stops at reconstruction; whether these representations help language modeling remains untested. We test two assumptions implicit in the optical-compression narrative: that vision-based compression provides unique advantages for text reconstruction from compressed representations, and that DeepSeek-OCR's reconstruction results are evidence that vision-based compression will be useful for language modeling. Comparing their vision encoder against simple alternatives--parameter-free mean pooling and a learned hierarchical encoder--we find that these simple approaches match or surpass vision for reconstruction at matched compression ratios, and outperform it for language modeling--where vision-based compression fails to beat truncation. The excitement around optical context compression outpaces the evidence. Code and checkpoints are available at https://github.com/ivnle/bad-autoencoding

Paperid: 270, https://arxiv.org/pdf/2512.03620.pdf GitHub

Abstract:
The protection of Intellectual Property (IP) in Large Language Models (LLMs) represents a critical challenge in contemporary AI research. While fingerprinting techniques have emerged as a fundamental mechanism for detecting unauthorized model usage, existing methods -- whether behavior-based or structural -- suffer from vulnerabilities such as false claim attacks or susceptible to weight manipulations. To overcome these limitations, we propose SELF, a novel intrinsic weight-based fingerprinting scheme that eliminates dependency on input and inherently resists false claims. SELF achieves robust IP protection through two key innovations: 1) unique, scalable and transformation-invariant fingerprint extraction via singular value and eigenvalue decomposition of LLM attention weights, and 2) effective neural network-based fingerprint similarity comparison based on few-shot learning and data augmentation. Experimental results demonstrate SELF maintains high IP infringement detection accuracy while showing strong robustness against various downstream modifications, including quantization, pruning, and fine-tuning attacks. Our code is available at https://github.com/HanxiuZhang/SELF_v2.

Paperid: 271, https://arxiv.org/pdf/2512.03619.pdf GitHub

Abstract:
Video generation has achieved remarkable progress in visual fidelity and controllability, enabling conditioning on text, layout, or motion. Among these, motion control - specifying object dynamics and camera trajectories - is essential for composing complex, cinematic scenes, yet existing interfaces remain limited. We introduce LAMP that leverages large language models (LLMs) as motion planners to translate natural language descriptions into explicit 3D trajectories for dynamic objects and (relatively defined) cameras. LAMP defines a motion domain-specific language (DSL), inspired by cinematography conventions. By harnessing program synthesis capabilities of LLMs, LAMP generates structured motion programs from natural language, which are deterministically mapped to 3D trajectories. We construct a large-scale procedural dataset pairing natural text descriptions with corresponding motion programs and 3D trajectories. Experiments demonstrate LAMP's improved performance in motion controllability and alignment with user intent compared to state-of-the-art alternatives establishing the first framework for generating both object and camera motions directly from natural language specifications. Code, models and data are available on our project page.

Paperid: 272, https://arxiv.org/pdf/2512.03601.pdf GitHub

Abstract:
Recent advancements in foundation models for 2D vision have substantially improved the analysis of dynamic scenes from monocular videos. However, despite their strong generalization capabilities, these models often lack 3D consistency, a fundamental requirement for understanding scene geometry and motion, thereby causing severe spatial misalignment and temporal flickering in complex 3D environments. In this paper, we present Motion4D, a novel framework that addresses these challenges by integrating 2D priors from foundation models into a unified 4D Gaussian Splatting representation. Our method features a two-part iterative optimization framework: 1) Sequential optimization, which updates motion and semantic fields in consecutive stages to maintain local consistency, and 2) Global optimization, which jointly refines all attributes for long-term coherence. To enhance motion accuracy, we introduce a 3D confidence map that dynamically adjusts the motion priors, and an adaptive resampling process that inserts new Gaussians into under-represented regions based on per-pixel RGB and semantic errors. Furthermore, we enhance semantic coherence through an iterative refinement process that resolves semantic inconsistencies by alternately optimizing the semantic fields and updating prompts of SAM2. Extensive evaluations demonstrate that our Motion4D significantly outperforms both 2D foundation models and existing 3D-based approaches across diverse scene understanding tasks, including point-based tracking, video object segmentation, and novel view synthesis. Our code is available at https://hrzhou2.github.io/motion4d-web/.

Paperid: 273, https://arxiv.org/pdf/2512.03597.pdf GitHub

Abstract:
Medical image segmentation is a cornerstone of modern clinical diagnostics. While Vision Transformers that leverage shifted window-based self-attention have established new benchmarks in this field, they are often hampered by a critical limitation: their localized attention mechanism struggles to effectively fuse local details with global context. This deficiency is particularly detrimental to challenging tasks such as the segmentation of microtumors and miniature organs, where both fine-grained boundary definition and broad contextual understanding are paramount. To address this gap, we propose HBFormer, a novel Hybrid-Bridge Transformer architecture. The 'Hybrid' design of HBFormer synergizes a classic U-shaped encoder-decoder framework with a powerful Swin Transformer backbone for robust hierarchical feature extraction. The core innovation lies in its 'Bridge' mechanism, a sophisticated nexus for multi-scale feature integration. This bridge is architecturally embodied by our novel Multi-Scale Feature Fusion (MFF) decoder. Departing from conventional symmetric designs, the MFF decoder is engineered to fuse multi-scale features from the encoder with global contextual information. It achieves this through a synergistic combination of channel and spatial attention modules, which are constructed from a series of dilated and depth-wise convolutions. These components work in concert to create a powerful feature bridge that explicitly captures long-range dependencies and refines object boundaries with exceptional precision. Comprehensive experiments on challenging medical image segmentation datasets, including multi-organ, liver tumor, and bladder tumor benchmarks, demonstrate that HBFormer achieves state-of-the-art results, showcasing its outstanding capabilities in microtumor and miniature organ segmentation. Code and models are available at: https://github.com/lzeeorno/HBFormer.

Paperid: 274, https://arxiv.org/pdf/2512.03577.pdf GitHub

Abstract:
Universal, transferable whole-slide image (WSI) representations are central to computational pathology. Incorporating multiple markers (e.g., immunohistochemistry, IHC) alongside H&E enriches H&E-based features with diverse, biologically meaningful information. However, progress is limited by the scarcity of well-aligned multi-stain datasets. Inter-stain misalignment shifts corresponding tissue across slides, hindering consistent patch-level features and degrading slide-level embeddings. To address this, we curated a slide-level aligned, five-stain dataset (H&E, HER2, KI67, ER, PGR) to enable paired H&E-IHC learning and robust cross-stain representation. Leveraging this dataset, we propose Cross-Stain Contrastive Learning (CSCL), a two-stage pretraining framework with a lightweight adapter trained using patch-wise contrastive alignment to improve the compatibility of H&E features with corresponding IHC-derived contextual cues, and slide-level representation learning with Multiple Instance Learning (MIL), which uses a cross-stain attention fusion module to integrate stain-specific patch features and a cross-stain global alignment module to enforce consistency among slide-level embeddings across different stains. Experiments on cancer subtype classification, IHC biomarker status classification, and survival prediction show consistent gains, yielding high-quality, transferable H&E slide-level representations. The code and data are available at https://github.com/lily-zyz/CSCL.

Paperid: 275, https://arxiv.org/pdf/2512.03558.pdf GitHub

Abstract:
The rise of Visual-Language Models (LVLMs) has unlocked new possibilities for seamlessly integrating visual and textual information. However, their ability to interpret cartographic maps remains largely unexplored. In this paper, we introduce CartoMapQA, a benchmark specifically designed to evaluate LVLMs' understanding of cartographic maps through question-answering tasks. The dataset includes over 2000 samples, each composed of a cartographic map, a question (with open-ended or multiple-choice answers), and a ground-truth answer. These tasks span key low-, mid- and high-level map interpretation skills, including symbol recognition, embedded information extraction, scale interpretation, and route-based reasoning. Our evaluation of both open-source and proprietary LVLMs reveals persistent challenges: models frequently struggle with map-specific semantics, exhibit limited geospatial reasoning, and are prone to Optical Character Recognition (OCR)-related errors. By isolating these weaknesses, CartoMapQA offers a valuable tool for guiding future improvements in LVLM architectures. Ultimately, it supports the development of models better equipped for real-world applications that depend on robust and reliable map understanding, such as navigation, geographic search, and urban planning. Our source code and data are openly available to the research community at: https://github.com/ungquanghuy-kddi/CartoMapQA.git

Paperid: 276, https://arxiv.org/pdf/2512.03534.pdf GitHub GitHub

Abstract:
Achieving precise alignment between user intent and generated visuals remains a central challenge in text-to-visual generation, as a single attempt often fails to produce the desired output. To handle this, prior approaches mainly scale the visual generation process (e.g., increasing sampling steps or seeds), but this quickly leads to a quality plateau. This limitation arises because the prompt, crucial for guiding generation, is kept fixed. To address this, we propose Prompt Redesign for Inference-time Scaling, coined PRIS, a framework that adaptively revises the prompt during inference in response to the scaled visual generations. The core idea of PRIS is to review the generated visuals, identify recurring failure patterns across visuals, and redesign the prompt accordingly before regenerating the visuals with the revised prompt. To provide precise alignment feedback for prompt revision, we introduce a new verifier, element-level factual correction, which evaluates the alignment between prompt attributes and generated visuals at a fine-grained level, achieving more accurate and interpretable assessments than holistic measures. Extensive experiments on both text-to-image and text-to-video benchmarks demonstrate the effectiveness of our approach, including a 15% gain on VBench 2.0. These results highlight that jointly scaling prompts and visuals is key to fully leveraging scaling laws at inference-time. Visualizations are available at the website: https://subin-kim-cv.github.io/PRIS.

Paperid: 277, https://arxiv.org/pdf/2512.03520.pdf GitHub

Abstract:
We present FloodDiffusion, a new framework for text-driven, streaming human motion generation. Given time-varying text prompts, FloodDiffusion generates text-aligned, seamless motion sequences with real-time latency. Unlike existing methods that rely on chunk-by-chunk or auto-regressive model with diffusion head, we adopt a diffusion forcing framework to model this time-series generation task under time-varying control events. We find that a straightforward implementation of vanilla diffusion forcing (as proposed for video models) fails to model real motion distributions. We demonstrate that to guarantee modeling the output distribution, the vanilla diffusion forcing must be tailored to: (i) train with a bi-directional attention instead of casual attention; (ii) implement a lower triangular time scheduler instead of a random one; (iii) utilize a continues time-varying way to introduce text conditioning. With these improvements, we demonstrate in the first time that the diffusion forcing-based framework achieves state-of-the-art performance on the streaming motion generation task, reaching an FID of 0.057 on the HumanML3D benchmark. Models, code, and weights are available. https://shandaai.github.io/FloodDiffusion/

Paperid: 278, https://arxiv.org/pdf/2512.03508.pdf GitHub

Abstract:
Recent domain generalized semantic segmentation (DGSS) studies have achieved notable improvements by distilling semantic knowledge from Vision-Language Models (VLMs). However, they overlook the semantic misalignment between visual and textual contexts, which arises due to the rigidity of a fixed context prompt learned on a single source domain. To this end, we present a novel domain generalization framework for semantic segmentation, namely Domain-aware Prompt-driven Masked Transformer (DPMFormer). Firstly, we introduce domain-aware prompt learning to facilitate semantic alignment between visual and textual cues. To capture various domain-specific properties with a single source dataset, we propose domain-aware contrastive learning along with the texture perturbation that diversifies the observable domains. Lastly, to establish a framework resilient against diverse environmental changes, we have proposed the domain-robust consistency learning which guides the model to minimize discrepancies of prediction from original and the augmented images. Through experiments and analyses, we demonstrate the superiority of the proposed framework, which establishes a new state-of-the-art on various DGSS benchmarks. The code is available at https://github.com/jone1222/DPMFormer.

Paperid: 279, https://arxiv.org/pdf/2512.03470.pdf GitHub

Abstract:
Infrared small target detection (ISTD) faces two major challenges: a lack of discernible target texture and severe background clutter, which results in the background obscuring the target. To enhance targets and suppress backgrounds, we propose the Basis Decomposition Module (BDM) as an extensible and lightweight module based on basis decomposition, which decomposes a complex feature into several basis features and enhances certain information while eliminating redundancy. Extending BDM leads to a series of modules, including the Spatial Difference Decomposition Module (SD$^\mathrm{2}$M), Spatial Difference Decomposition Downsampling Module (SD$^\mathrm{3}$M), and Temporal Difference Decomposition Module (TD$^\mathrm{2}$M). Based on these modules, we develop the Spatial Difference Decomposition Network (SD$^\mathrm{2}$Net) for single-frame ISTD (SISTD) and the Spatiotemporal Difference Decomposition Network (STD$^\mathrm{2}$Net) for multi-frame ISTD (MISTD). SD$^\mathrm{2}$Net integrates SD$^\mathrm{2}$M and SD$^\mathrm{3}$M within an adapted U-shaped architecture. We employ TD$^\mathrm{2}$M to introduce motion information, which transforms SD$^\mathrm{2}$Net into STD$^\mathrm{2}$Net. Extensive experiments on SISTD and MISTD datasets demonstrate state-of-the-art (SOTA) performance. On the SISTD task, SD$^\mathrm{2}$Net performs well compared to most established networks. On the MISTD datasets, STD$^\mathrm{2}$Net achieves a mIoU of 87.68\%, outperforming SD$^\mathrm{2}$Net, which achieves a mIoU of 64.97\%. Our codes are available: https://github.com/greekinRoma/IRSTD_HC_Platform.

Paperid: 280, https://arxiv.org/pdf/2512.03453.pdf GitHub

Abstract:
Recent advances in video generation have enabled the synthesis of high-quality and visually realistic clips using diffusion transformer models. However, most existing approaches operate purely in the 2D pixel space and lack explicit mechanisms for modeling 3D structures, often resulting in temporally inconsistent geometries, implausible motions, and structural artifacts. In this work, we introduce geometric regularization losses into video generation by augmenting latent diffusion models with per-frame depth prediction. We adopted depth as the geometric representation because of the great progress in depth prediction and its compatibility with image-based latent encoders. Specifically, to enforce structural consistency over time, we propose a multi-view geometric loss that aligns the predicted depth maps across frames within a shared 3D coordinate system. Our method bridges the gap between appearance generation and 3D structure modeling, leading to improved spatio-temporal coherence, shape consistency, and physical plausibility. Experiments across multiple datasets show that our approach produces significantly more stable and geometrically consistent results than existing baselines.

Paperid: 281, https://arxiv.org/pdf/2512.03445.pdf GitHub

Abstract:
Vision-language pretraining (VLP) has emerged as a powerful paradigm in medical image analysis, enabling representation learning from large-scale image-text pairs without relying on expensive manual annotations. However, existing methods often struggle with the noise inherent in web-collected data and the complexity of unstructured long medical texts. To address these challenges, we propose a novel VLP framework integrating a Multi-Agent data GENeration (MAGEN) system and Ontology-based Multi-Aspect Knowledge-Enhanced (O-MAKE) pretraining. First, MAGEN enhances data quality by synthesizing knowledge-enriched descriptions via a foundation model-assisted captioning and retrieval-based verification pipeline. Second, O-MAKE addresses the difficulty of learning from long, unstructured texts by decomposing them into distinct knowledge aspects. This facilitates fine-grained alignment at both global and patch levels, while explicitly modeling medical concept relationships through ontology-guided mechanisms. We validate our framework in the field of dermatology, where comprehensive experiments demonstrate the effectiveness of each component. Our approach achieves state-of-the-art zero-shot performance on disease classification and cross-modal retrieval tasks across eight datasets. Our code and the augmented dataset Derm1M-AgentAug, comprising over 400k skin-image-text pairs, will be released at https://github.com/SiyuanYan1/Derm1M.

Paperid: 282, https://arxiv.org/pdf/2512.03397.pdf GitHub

Abstract:
LiDAR-inertial odometry (LIO) is an active research area, as it enables accurate real-time state estimation in GPS-denied environments. Recent advances in map data structures and spatial indexing have significantly improved the efficiency of LIO systems. Nevertheless, we observe that two aspects may still leave room for improvement: (1) nearest neighbor search often requires examining multiple spatial units to gather sufficient points for plane fitting, and (2) plane parameters are typically recomputed at every iteration despite unchanged map geometry. Motivated by these observations, we propose Surfel-LIO, which employs a hierarchical voxel structure (hVox) with pre-computed surfel representation. This design enables O(1) correspondence retrieval without runtime neighbor enumeration or plane fitting, combined with Z-order curve encoding for cache-friendly spatial indexing. Experimental results on the M3DGR dataset demonstrate that our method achieves significantly faster processing speed compared to recent state-of-the-art methods while maintaining comparable state estimation accuracy. Our implementation is publicly available at https://github.com/93won/lidar_inertial_odometry.

Paperid: 283, https://arxiv.org/pdf/2512.03383.pdf GitHub

Abstract:
Deploying large language models (LLMs) on mobile platforms faces significant challenges due to the limited memory and shared computational resources of the device. Resource availability may be an issue as it is directly impacted by the current device workload, adding to the uncertainty of model deployment. We introduce UniQL, a unified post-training quantization and low-rank compression framework with on-device configurable pruning rates for edge LLMs. UniQL is a general framework that integrates quantization and low-rank compression for Transformers, State Space Models (SSMs), and hybrid models to support diverse edge applications. In our proposed joint framework, we introduce an efficient structured weight-sorting method that speeds up computation by 20x, quantization-aware singular value decomposition (SVD) to minimize quantization errors, state-aware weight sorting for SSMs, and a fused rotary positional embedding (RoPE) kernel for pruned models. Our framework performs weight-sorting, fine-tuning, and quantization in the cloud in a single-pass workflow, while enabling on-device configurable pruning rates up to 35%. Our experiments show that quantized and pruned models achieve a memory reduction of 4x-5.7x and a token-throughput improvement of 2.7x-3.4x, maintaining accuracy within 5% of the original models at 15% pruning across Transformers (Llama3 and Qwen2.5), SSMs (Mamba2), and hybrid models (Nemotron-H and Bamba-v2). The code and quantized models are available at: https://github.com/enyac-group/UniQL.

Paperid: 284, https://arxiv.org/pdf/2512.03370.pdf GitHub

Abstract:
We introduce ShelfGaussian, an open-vocabulary multi-modal Gaussian-based 3D scene understanding framework supervised by off-the-shelf vision foundation models (VFMs). Gaussian-based methods have demonstrated superior performance and computational efficiency across a wide range of scene understanding tasks. However, existing methods either model objects as closed-set semantic Gaussians supervised by annotated 3D labels, neglecting their rendering ability, or learn open-set Gaussian representations via purely 2D self-supervision, leading to degraded geometry and limited to camera-only settings. To fully exploit the potential of Gaussians, we propose a Multi-Modal Gaussian Transformer that enables Gaussians to query features from diverse sensor modalities, and a Shelf-Supervised Learning Paradigm that efficiently optimizes Gaussians with VFM features jointly at 2D image and 3D scene levels. We evaluate ShelfGaussian on various perception and planning tasks. Experiments on Occ3D-nuScenes demonstrate its state-of-the-art zero-shot semantic occupancy prediction performance. ShelfGaussian is further evaluated on an unmanned ground vehicle (UGV) to assess its in the-wild performance across diverse urban scenarios. Project website: https://lunarlab-gatech.github.io/ShelfGaussian/.

Paperid: 285, https://arxiv.org/pdf/2512.03369.pdf GitHub

Abstract:
Fine-grained wildfire spread prediction is crucial for enhancing emergency response efficacy and decision-making precision. However, existing research predominantly focuses on coarse spatiotemporal scales and relies on low-resolution satellite data, capturing only macroscopic fire states while fundamentally constraining high-precision localized fire dynamics modeling capabilities. To bridge this gap, we present FireSentry, a provincial-scale multi-modal wildfire dataset characterized by sub-meter spatial and sub-second temporal resolution. Collected using synchronized UAV platforms, FireSentry provides visible and infrared video streams, in-situ environmental measurements, and manually validated fire masks. Building on FireSentry, we establish a comprehensive benchmark encompassing physics-based, data-driven, and generative models, revealing the limitations of existing mask-only approaches. Our analysis proposes FiReDiff, a novel dual-modality paradigm that first predicts future video sequences in the infrared modality, and then precisely segments fire masks in the mask modality based on the generated dynamics. FiReDiff achieves state-of-the-art performance, with video quality gains of 39.2% in PSNR, 36.1% in SSIM, 50.0% in LPIPS, 29.4% in FVD, and mask accuracy gains of 3.3% in AUPRC, 59.1% in F1 score, 42.9% in IoU, and 62.5% in MSE when applied to generative models. The FireSentry benchmark dataset and FiReDiff paradigm collectively advance fine-grained wildfire forecasting and dynamic disaster simulation. The processed benchmark dataset is publicly available at: https://github.com/Munan222/FireSentry-Benchmark-Dataset.

Paperid: 286, https://arxiv.org/pdf/2512.03343.pdf GitHub

Abstract:
Autoregressive Language Models (LLMs) trained on Next-Token Prediction (NTP) often suffer from ``Topic Drift'' where the generation wanders away from the initial prompt due to a reliance on local associations rather than global planning \citep{holtzman2019curious}. While scaling model size mitigates this \citep{brown2020language}, the fundamental myopia of the NTP objective remains. In this work, we introduce the Idea-Gated Transformer, a novel architecture that separates semantic planning from syntactic generation. We introduce an auxiliary ``Idea Head'' trained to predict the bag-of-words distribution for a future context window, creating a latent ``Concept Vector'' that actively gates the main vocabulary during generation. We propose a differentiable gating mechanism that suppresses semantically irrelevant tokens, effectively pruning the search space in real-time. Experiments on WikiText-103 demonstrate that while the Idea-Gated model achieves comparable validation perplexity to a standard GPT-2 baseline, it exhibits significantly superior Domain Retention. Qualitative and quantitative analysis reveals that the gating mechanism successfully locks generation into specific semantic clusters (e.g., Finance, Science) and resists associative drift, offering a parameter-efficient path toward more controllable language modeling.

Paperid: 287, https://arxiv.org/pdf/2512.03339.pdf GitHub

Abstract:
Ejection fraction (EF) is a crucial metric for assessing cardiac function and diagnosing conditions such as heart failure. Traditionally, EF estimation requires manual tracing and domain expertise, making the process time-consuming and subject to interobserver variability. Most current deep learning methods for EF prediction are black-box models with limited transparency, which reduces clinical trust. Some post-hoc explainability methods have been proposed to interpret the decision-making process after the prediction is made. However, these explanations do not guide the model's internal reasoning and therefore offer limited reliability in clinical applications. To address this, we introduce ProtoEFNet, a novel video-based prototype learning model for continuous EF regression. The model learns dynamic spatiotemporal prototypes that capture clinically meaningful cardiac motion patterns. Additionally, the proposed Prototype Angular Separation (PAS) loss enforces discriminative representations across the continuous EF spectrum. Our experiments on the EchonetDynamic dataset show that ProtoEFNet can achieve accuracy on par with its non-interpretable counterpart while providing clinically relevant insight. The ablation study shows that the proposed loss boosts performance with a 2% increase in F1 score from 77.67$\pm$2.68 to 79.64$\pm$2.10. Our source code is available at: https://github.com/DeepRCL/ProtoEF

Paperid: 288, https://arxiv.org/pdf/2512.03317.pdf GitHub

Abstract:
Accurate environmental representations are essential for autonomous driving, providing the foundation for safe and efficient navigation. Traditionally, high-definition (HD) maps are providing this representation of the static road infrastructure to the autonomous system a priori. However, because the real world is constantly changing, such maps must be constructed online from on-board sensor data. Navigation-grade standard-definition (SD) maps are widely available, but their resolution is insufficient for direct deployment. Instead, they can be used as coarse prior to guide the online map construction process. We propose NavMapFusion, a diffusion-based framework that performs iterative denoising conditioned on high-fidelity sensor data and on low-fidelity navigation maps. This paper strives to answer: (1) How can coarse, potentially outdated navigation maps guide online map construction? (2) What advantages do diffusion models offer for map fusion? We demonstrate that diffusion-based map construction provides a robust framework for map fusion. Our key insight is that discrepancies between the prior map and online perception naturally correspond to noise within the diffusion process; consistent regions reinforce the map construction, whereas outdated segments are suppressed. On the nuScenes benchmark, NavMapFusion conditioned on coarse road lines from OpenStreetMap data reaches a 21.4% relative improvement on 100 m, and even stronger improvements on larger perception ranges, while maintaining real-time capabilities. By fusing low-fidelity priors with high-fidelity sensor data, the proposed method generates accurate and up-to-date environment representations, guiding towards safer and more reliable autonomous driving. The code is available at https://github.com/tmonnin/navmapfusion

Paperid: 289, https://arxiv.org/pdf/2512.03312.pdf GitHub

Abstract:
The function of biomolecules such as proteins depends on their ability to interconvert between a wide range of structures or "conformations." Researchers have endeavored for decades to develop computational methods to predict the distribution of conformations, which is far harder to determine experimentally than a static folded structure. We present ConforMix, an inference-time algorithm that enhances sampling of conformational distributions using a combination of classifier guidance, filtering, and free energy estimation. Our approach upgrades diffusion models -- whether trained for static structure prediction or conformational generation -- to enable more efficient discovery of conformational variability without requiring prior knowledge of major degrees of freedom. ConforMix is orthogonal to improvements in model pretraining and would benefit even a hypothetical model that perfectly reproduced the Boltzmann distribution. Remarkably, when applied to a diffusion model trained for static structure prediction, ConforMix captures structural changes including domain motion, cryptic pocket flexibility, and transporter cycling, while avoiding unphysical states. Case studies of biologically critical proteins demonstrate the scalability, accuracy, and utility of this method.

Paperid: 290, https://arxiv.org/pdf/2512.03216.pdf GitHub

Abstract:
Scintillators are transparent materials that interact with high-energy particles and emit visible light as a result. They are used in state of the art methods of measuring high-energy particles and radiation sources. Most existing methods use fast single-pixel detectors to detect and time scintillation events. Cameras provide spatial resolution but can only capture an average over many events, making it difficult to image the events associated with an individual particle. Emerging single-photon avalanche diode cameras combine speed and spatial resolution to enable capturing images of individual events. This allows us to use machine vision techniques to analyze events, enabling new types of detectors. The main challenge is the very low brightness of the events. Techniques have to work with a very limited number of photons. We propose a kaleidoscopic scintillator to increase light collection in a single-photon camera while preserving the event's spatial information. The kaleidoscopic geometry creates mirror reflections of the event in known locations for a given event location that are captured by the camera. We introduce theory for imaging an event in a kaleidoscopic scintillator and an algorithm to estimate the event's 3D position. We find that the kaleidoscopic scintillator design provides sufficient light collection to perform high-resolution event measurements for advanced radiation imaging techniques using a commercial CMOS single-photon camera. Code and data are available at https://github.com/bocchs/kaleidoscopic_scintillator.

Paperid: 291, https://arxiv.org/pdf/2512.03195.pdf GitHub

Abstract:
This study investigates the potential of language models to improve the classification of labor market information by linking job vacancy texts to two major European frameworks: the European Skills, Competences, Qualifications and Occupations (ESCO) taxonomy and the European Qualifications Framework (EQF). We examine and compare two prominent methodologies from the literature: Sentence Linking and Entity Linking. In support of ongoing research, we release an open-source tool, incorporating these two methodologies, designed to facilitate further work on labor classification and employment discourse. To move beyond surface-level skill extraction, we introduce two annotated datasets specifically aimed at evaluating how occupations and qualifications are represented within job vacancy texts. Additionally, we examine different ways to utilize generative large language models for this task. Our findings contribute to advancing the state of the art in job entity extraction and offer computational infrastructure for examining work, skills, and labor market narratives in a digitally mediated economy. Our code is made publicly available: https://github.com/tabiya-tech/tabiya-livelihoods-classifier

Paperid: 292, https://arxiv.org/pdf/2512.03175.pdf GitHub

Abstract:
The Seifert-van Kampen theorem computes the fundamental group of a space from the fundamental groups of its constituents. We formalize this theorem within the framework of computational paths -- an approach to equality where witnesses are explicit sequences of rewrites governed by the confluent, terminating LND_EQ-TRS. Our contributions are: (i) pushouts as higher-inductive types with explicit path constructors; (ii) free products and amalgamated free products as quotients of word representations; (iii) an encode-decode proof establishing pi_1(Pushout(A,B,C)) = pi_1(A) *_{pi_1(C)} pi_1(B); and (iv) applications to the figure-eight (pi_1(S^1 V S^1) = Z * Z), 2-sphere (pi_1(S^2) = 1), and 3-sphere (pi_1(S^3) = 1 with Hopf fibration context). Recent extensions include: higher homotopy groups pi_n via the weak omega-groupoid structure (with pi_2 abelian via Eckmann-Hilton, and pi_2 = 1 in the 1-groupoid truncated setting); truncation levels connecting the framework to HoTT n-types; automated path simplification tactics; basic covering space theory with pi_1-actions on fibers; fibration theory with long exact sequences; and Eilenberg-MacLane space characterization (S^1 = K(Z,1)). The framework makes coherence witnesses explicit as rewrite derivations. The development is formalized in Lean 4 with over 25,000 lines of mechanized proofs. This demonstrates that the encode-decode method for higher-inductive types becomes fully constructive when path equality is decidable via normalization.

Paperid: 293, https://arxiv.org/pdf/2512.03127.pdf GitHub

Abstract:
Nuclear Magnetic Resonance (NMR) spectroscopy is a cornerstone technique for determining the structures of small molecules and is especially critical in the discovery of novel natural products and clinical therapeutics. Yet, interpreting NMR spectra remains a time-consuming, manual process requiring extensive domain expertise. We introduce ChefNMR (CHemical Elucidation From NMR), an end-to-end framework that directly predicts an unknown molecule's structure solely from its 1D NMR spectra and chemical formula. We frame structure elucidation as conditional generation from an atomic diffusion model built on a non-equivariant transformer architecture. To model the complex chemical groups found in natural products, we generated a dataset of simulated 1D NMR spectra for over 111,000 natural products. ChefNMR predicts the structures of challenging natural product compounds with an unsurpassed accuracy of over 65%. This work takes a significant step toward solving the grand challenge of automating small-molecule structure elucidation and highlights the potential of deep learning in accelerating molecular discovery. Code is available at https://github.com/ml-struct-bio/chefnmr.

Paperid: 294, https://arxiv.org/pdf/2512.03125.pdf GitHub

Abstract:
Unified Multimodal Generative Models (UMGMs) unify visual understanding and image generation within a single autoregressive framework. However, their ability to continually learn new tasks is severely hindered by catastrophic forgetting, both within a modality (intra-modal) and across modalities (inter-modal). While intra-modal forgetting has been studied in prior continual learning (CL) work, inter-modal forgetting remains largely unexplored. In this paper, we identify and empirically validate this phenomenon in UMGMs and provide a theoretical explanation rooted in gradient conflict between modalities. To address both intra- and inter-modal forgetting, we propose Modality-Decoupled Experts (MoDE), a lightweight and scalable architecture that isolates modality-specific updates to mitigate the gradient conflict and leverages knowledge distillation to prevent catastrophic forgetting and preserve pre-trained capabilities. Unlike previous CL methods that remain modality-coupled and suffer from modality gradient conflict, MoDE explicitly decouples modalities to prevent interference. Experiments across diverse benchmarks demonstrate that MoDE significantly mitigates both inter- and intra-modal forgetting, outperforming prior CL baselines in unified multimodal generation settings. Codes will be publicly available: https://github.com/Christina200/MoDE-official.git

Paperid: 295, https://arxiv.org/pdf/2512.03111.pdf GitHub

Abstract:
Single-cell RNA sequencing (scRNA-seq) is essential for decoding tumor heterogeneity. However, pan-cancer research still faces two key challenges: learning discriminative and efficient single-cell representations, and establishing a comprehensive evaluation benchmark. In this paper, we introduce PanFoMa, a lightweight hybrid neural network that combines the strengths of Transformers and state-space models to achieve a balance between performance and efficiency. PanFoMa consists of a front-end local-context encoder with shared self-attention layers to capture complex, order-independent gene interactions; and a back-end global sequential feature decoder that efficiently integrates global context using a linear-time state-space model. This modular design preserves the expressive power of Transformers while leveraging the scalability of Mamba to enable transcriptome modeling, effectively capturing both local and global regulatory signals. To enable robust evaluation, we also construct a large-scale pan-cancer single-cell benchmark, PanFoMaBench, containing over 3.5 million high-quality cells across 33 cancer subtypes, curated through a rigorous preprocessing pipeline. Experimental results show that PanFoMa outperforms state-of-the-art models on our pan-cancer benchmark (+4.0\%) and across multiple public tasks, including cell type annotation (+7.4\%), batch integration (+4.0\%) and multi-omics integration (+3.1\%). The code is available at https://github.com/Xiaoshui-Huang/PanFoMa.

Paperid: 296, https://arxiv.org/pdf/2512.03099.pdf GitHub

Abstract:
Graph Neural Networks (GNNs) have become indispensable in critical domains such as drug discovery, social network analysis, and recommendation systems, yet their black-box nature hinders deployment in scenarios requiring transparency and accountability. While Shapley value-based methods offer mathematically principled explanations by quantifying each component's contribution to predictions, computing exact values requires evaluating $2^n$ coalitions (or aggregating over $n!$ permutations), which is intractable for real-world graphs. Existing approximation strategies sacrifice either fidelity or efficiency, limiting their practical utility. We introduce QGShap, a quantum computing approach that leverages amplitude amplification to achieve quadratic speedups in coalition evaluation while maintaining exact Shapley computation. Unlike classical sampling or surrogate methods, our approach provides fully faithful explanations without approximation trade-offs for tractable graph sizes. We conduct empirical evaluations on synthetic graph datasets, demonstrating that QGShap achieves consistently high fidelity and explanation accuracy, matching or exceeding the performance of classical methods across all evaluation metrics. These results collectively demonstrate that QGShap not only preserves exact Shapley faithfulness but also delivers interpretable, stable, and structurally consistent explanations that align with the underlying graph reasoning of GNNs. The implementation of QGShap is available at https://github.com/smlab-niser/qgshap.

Paperid: 297, https://arxiv.org/pdf/2512.03074.pdf GitHub

Abstract:
Graph Neural Networks (GNNs) have demonstrated exceptional efficacy in relational learning tasks, including node classification and link prediction. However, their application raises significant fairness concerns, as GNNs can perpetuate and even amplify societal biases against protected groups defined by sensitive attributes such as race or gender. These biases are often inherent in the node features, structural topology, and message-passing mechanisms of the graph itself. A critical limitation of existing fairness-aware GNN methods is their reliance on the strong assumption that sensitive attributes are fully available for all nodes during training--a condition that poses a practical impediment due to privacy concerns and data collection constraints. To address this gap, we propose a novel, model-agnostic fairness regularization framework designed for the realistic scenario where sensitive attributes are only partially available. Our approach formalizes a fairness-aware objective function that integrates both equal opportunity and statistical parity as differentiable regularization terms. Through a comprehensive empirical evaluation across five real-world benchmark datasets, we demonstrate that the proposed method significantly mitigates bias across key fairness metrics while maintaining competitive node classification performance. Results show that our framework consistently outperforms baseline models in achieving a favorable fairness-accuracy trade-off, with minimal degradation in predictive accuracy. The datasets and source code will be publicly released at https://github.com/mtavassoli/GNN-FC.

Paperid: 298, https://arxiv.org/pdf/2512.03056.pdf GitHub

Abstract:
Diffusion models like Stable Diffusion (SD) drive a vibrant open-source ecosystem including fully fine-tuned checkpoints and parameter-efficient adapters such as LoRA, LyCORIS, and ControlNet. However, these adaptation components are tightly coupled to a specific base model, making them difficult to reuse when the base model is upgraded (e.g., from SD 1.x to 2.x) due to substantial changes in model parameters and architecture. In this work, we propose Delta Sampling (DS), a novel method that enables knowledge transfer across base models with different architectures, without requiring access to the original training data. DS operates entirely at inference time by leveraging the delta: the difference in model predictions before and after the adaptation of a base model. This delta is then used to guide the denoising process of a new base model. We evaluate DS across various SD versions, demonstrating that DS achieves consistent improvements in creating desired effects (e.g., visual styles, semantic concepts, and structures) under different sampling strategies. These results highlight DS as an effective, plug-and-play mechanism for knowledge transfer in diffusion-based image synthesis. Code:~ https://github.com/Zhidong-Gao/DeltaSampling

Paperid: 299, https://arxiv.org/pdf/2512.03047.pdf GitHub

Abstract:
Large language model safety is usually assessed with static benchmarks, but key failures are dynamic: value drift under distribution shift, jailbreak attacks, and slow degradation of alignment in deployment. Building on a recent Second Law of Intelligence that treats ethical entropy as a state variable which tends to increase unless countered by alignment work, we make this framework operational for large language models. We define a five-way behavioral taxonomy, train a classifier to estimate ethical entropy S(t) from model transcripts, and measure entropy dynamics for base and instruction-tuned variants of four frontier models across stress tests. Base models show sustained entropy growth, while tuned variants suppress drift and reduce ethical entropy by roughly eighty percent. From these trajectories we estimate an effective alignment work rate gamma_eff and embed S(t) and gamma_eff in a monitoring pipeline that raises alerts when entropy drift exceeds a stability threshold, enabling run-time oversight of value drift.

Paperid: 300, https://arxiv.org/pdf/2512.03045.pdf GitHub

Abstract:
Multi-view diffusion models have recently emerged as a powerful paradigm for novel view synthesis, yet the underlying mechanism that enables their view-consistency remains unclear. In this work, we first verify that the attention maps of these models acquire geometric correspondence throughout training, attending to the geometrically corresponding regions across reference and target views for view-consistent generation. However, this correspondence signal remains incomplete, with its accuracy degrading under large viewpoint changes. Building on these findings, we introduce CAMEO, a simple yet effective training technique that directly supervises attention maps using geometric correspondence to enhance both the training efficiency and generation quality of multi-view diffusion models. Notably, supervising a single attention layer is sufficient to guide the model toward learning precise correspondences, thereby preserving the geometry and structure of reference images, accelerating convergence, and improving novel view synthesis performance. CAMEO reduces the number of training iterations required for convergence by half while achieving superior performance at the same iteration counts. We further demonstrate that CAMEO is model-agnostic and can be applied to any multi-view diffusion model.

Paperid: 301, https://arxiv.org/pdf/2512.03043.pdf GitHub

Abstract:
Reinforcement learning (RL) has recently achieved remarkable success in eliciting visual reasoning within Multimodal Large Language Models (MLLMs). However, existing approaches typically train separate models for different tasks and treat image and video reasoning as disjoint domains. This results in limited scalability toward a multimodal reasoning generalist, which restricts practical versatility and hinders potential knowledge sharing across tasks and modalities. To this end, we propose OneThinker, an all-in-one reasoning model that unifies image and video understanding across diverse fundamental visual tasks, including question answering, captioning, spatial and temporal grounding, tracking, and segmentation. To achieve this, we construct the OneThinker-600k training corpus covering all these tasks and employ commercial models for CoT annotation, resulting in OneThinker-SFT-340k for SFT cold start. Furthermore, we propose EMA-GRPO to handle reward heterogeneity in multi-task RL by tracking task-wise moving averages of reward standard deviations for balanced optimization. Extensive experiments on diverse visual benchmarks show that OneThinker delivers strong performance on 31 benchmarks, across 10 fundamental visual understanding tasks. Moreover, it exhibits effective knowledge transfer between certain tasks and preliminary zero-shot generalization ability, marking a step toward a unified multimodal reasoning generalist. All code, model, and data are released.

Paperid: 302, https://arxiv.org/pdf/2512.03042.pdf GitHub

Abstract:
We introduce PPTArena, a benchmark for PowerPoint editing that measures reliable modifications to real slides under natural-language instructions. In contrast to image-PDF renderings or text-to-slide generation, PPTArena focuses on in-place editing across 100 decks, 2125 slides, and over 800 targeted edits covering text, charts, tables, animations, and master-level styles. Each case includes a ground-truth deck, a fully specified target outcome, and a dual VLM-as-judge pipeline that separately scores instruction following and visual quality using both structural diffs and slide images. Building on this setting, we propose PPTPilot, a structure-aware slide-editing agent that plans semantic edit sequences, routes between high-level programmatic tools and deterministic XML operations for precise control, and verifies outputs through an iterative plan-edit-check loop against task-specific constraints. In our experiments, PPTPilot outperforms strong proprietary agents and frontier VLM systems by over 10 percentage points on compound, layout-sensitive, and cross-slide edits, with particularly large gains in visual fidelity and deck-wide consistency. Despite these improvements, existing agents still underperform on long-horizon, document-scale tasks in PPTArena, highlighting the remaining challenges in reliable PPT editing.

Paperid: 303, https://arxiv.org/pdf/2512.03041.pdf GitHub

Abstract:
Current video generation techniques excel at single-shot clips but struggle to produce narrative multi-shot videos, which require flexible shot arrangement, coherent narrative, and controllability beyond text prompts. To tackle these challenges, we propose MultiShotMaster, a framework for highly controllable multi-shot video generation. We extend a pretrained single-shot model by integrating two novel variants of RoPE. First, we introduce Multi-Shot Narrative RoPE, which applies explicit phase shift at shot transitions, enabling flexible shot arrangement while preserving the temporal narrative order. Second, we design Spatiotemporal Position-Aware RoPE to incorporate reference tokens and grounding signals, enabling spatiotemporal-grounded reference injection. In addition, to overcome data scarcity, we establish an automated data annotation pipeline to extract multi-shot videos, captions, cross-shot grounding signals and reference images. Our framework leverages the intrinsic architectural properties to support multi-shot video generation, featuring text-driven inter-shot consistency, customized subject with motion control, and background-driven customized scene. Both shot count and duration are flexibly configurable. Extensive experiments demonstrate the superior performance and outstanding controllability of our framework.

Paperid: 304, https://arxiv.org/pdf/2512.03040.pdf GitHub

Abstract:
We investigate whether video generative models can exhibit visuospatial intelligence, a capability central to human cognition, using only visual data. To this end, we present Video4Spatial, a framework showing that video diffusion models conditioned solely on video-based scene context can perform complex spatial tasks. We validate on two tasks: scene navigation - following camera-pose instructions while remaining consistent with 3D geometry of the scene, and object grounding - which requires semantic localization, instruction following, and planning. Both tasks use video-only inputs, without auxiliary modalities such as depth or poses. With simple yet effective design choices in the framework and data curation, Video4Spatial demonstrates strong spatial understanding from video context: it plans navigation and grounds target objects end-to-end, follows camera-pose instructions while maintaining spatial consistency, and generalizes to long contexts and out-of-domain environments. Taken together, these results advance video generative models toward general visuospatial reasoning.

Paperid: 305, https://arxiv.org/pdf/2512.03036.pdf GitHub

Abstract:
Despite progress in video-to-audio generation, the field focuses predominantly on mono output, lacking spatial immersion. Existing binaural approaches remain constrained by a two-stage pipeline that first generates mono audio and then performs spatialization, often resulting in error accumulation and spatio-temporal inconsistencies. To address this limitation, we introduce the task of end-to-end binaural spatial audio generation directly from silent video. To support this task, we present the BiAudio dataset, comprising approximately 97K video-binaural audio pairs spanning diverse real-world scenes and camera rotation trajectories, constructed through a semi-automated pipeline. Furthermore, we propose ViSAudio, an end-to-end framework that employs conditional flow matching with a dual-branch audio generation architecture, where two dedicated branches model the audio latent flows. Integrated with a conditional spacetime module, it balances consistency between channels while preserving distinctive spatial characteristics, ensuring precise spatio-temporal alignment between audio and the input video. Comprehensive experiments demonstrate that ViSAudio outperforms existing state-of-the-art methods across both objective metrics and subjective evaluations, generating high-quality binaural audio with spatial immersion that adapts effectively to viewpoint changes, sound-source motion, and diverse acoustic environments. Project website: https://kszpxxzmc.github.io/ViSAudio-project.

Paperid: 306, https://arxiv.org/pdf/2512.03034.pdf GitHub

Abstract:
We propose MAViD, a novel Multimodal framework for Audio-Visual Dialogue understanding and generation. Existing approaches primarily focus on non-interactive systems and are limited to producing constrained and unnatural human speech.The primary challenge of this task lies in effectively integrating understanding and generation capabilities, as well as achieving seamless multimodal audio-video fusion. To solve these problems, we propose a Conductor-Creator architecture that divides the dialogue system into two primary components.The Conductor is tasked with understanding, reasoning, and generating instructions by breaking them down into motion and speech components, thereby enabling fine-grained control over interactions. The Creator then delivers interactive responses based on these instructions.Furthermore, to address the difficulty of generating long videos with consistent identity, timbre, and tone using dual DiT structures, the Creator adopts a structure that combines autoregressive (AR) and diffusion models. The AR model is responsible for audio generation, while the diffusion model ensures high-quality video generation.Additionally, we propose a novel fusion module to enhance connections between contextually consecutive clips and modalities, enabling synchronized long-duration audio-visual content generation.Extensive experiments demonstrate that our framework can generate vivid and contextually coherent long-duration dialogue interactions and accurately interpret users' multimodal queries.

Paperid: 307, https://arxiv.org/pdf/2512.03018.pdf GitHub

Abstract:
The boundary representation (B-Rep) is the standard data structure used in Computer-Aided Design (CAD) for defining solid models. Despite recent progress, directly generating B-Reps end-to-end with precise geometry and watertight topology remains a challenge. This paper presents AutoBrep, a novel Transformer model that autoregressively generates B-Reps with high quality and validity. AutoBrep employs a unified tokenization scheme that encodes both geometric and topological characteristics of a B-Rep model as a sequence of discrete tokens. Geometric primitives (i.e., surfaces and curves) are encoded as latent geometry tokens, and their structural relationships are defined as special topological reference tokens. Sequence order in AutoBrep naturally follows a breadth first traversal of the B-Rep face adjacency graph. At inference time, neighboring faces and edges along with their topological structure are progressively generated. Extensive experiments demonstrate the advantages of our unified representation when coupled with next-token prediction for B-Rep generation. AutoBrep outperforms baselines with better quality and watertightness. It is also highly scalable to complex solids with good fidelity and inference speed. We further show that autocompleting B-Reps is natively supported through our unified tokenization, enabling user-controllable CAD generation with minimal changes. Code is available at https://github.com/AutodeskAILab/AutoBrep.

Paperid: 308, https://arxiv.org/pdf/2512.03013.pdf GitHub

Abstract:
Editing portrait videos is a challenging task that requires flexible yet precise control over a wide range of modifications, such as appearance changes, expression edits, or the addition of objects. The key difficulty lies in preserving the subject's original temporal behavior, demanding that every edited frame remains precisely synchronized with the corresponding source frame. We present Sync-LoRA, a method for editing portrait videos that achieves high-quality visual modifications while maintaining frame-accurate synchronization and identity consistency. Our approach uses an image-to-video diffusion model, where the edit is defined by modifying the first frame and then propagated to the entire sequence. To enable accurate synchronization, we train an in-context LoRA using paired videos that depict identical motion trajectories but differ in appearance. These pairs are automatically generated and curated through a synchronization-based filtering process that selects only the most temporally aligned examples for training. This training setup teaches the model to combine motion cues from the source video with the visual changes introduced in the edited first frame. Trained on a compact, highly curated set of synchronized human portraits, Sync-LoRA generalizes to unseen identities and diverse edits (e.g., modifying appearance, adding objects, or changing backgrounds), robustly handling variations in pose and expression. Our results demonstrate high visual fidelity and strong temporal coherence, achieving a robust balance between edit fidelity and precise motion preservation.

Paperid: 309, https://arxiv.org/pdf/2512.03010.pdf GitHub

Abstract:
LiDAR-captured point clouds are often considered the gold standard in active 3D reconstruction. While their accuracy is exceptional in flat regions, the capturing is susceptible to miss small geometric structures and may fail with dark, absorbent materials. Alternatively, capturing multiple photos of the scene and applying 3D photogrammetry can infer these details as they often represent feature-rich regions. However, the accuracy of LiDAR for featureless regions is rarely reached. Therefore, we suggest combining the strengths of LiDAR and camera-based capture by introducing SurfFill: a Gaussian surfel-based LiDAR completion scheme. We analyze LiDAR capturings and attribute LiDAR beam divergence as a main factor for artifacts, manifesting mostly at thin structures and edges. We use this insight to introduce an ambiguity heuristic for completed scans by evaluating the change in density in the point cloud. This allows us to identify points close to missed areas, which we can then use to grow additional points from to complete the scan. For this point growing, we constrain Gaussian surfel reconstruction [Huang et al. 2024] to focus optimization and densification on these ambiguous areas. Finally, Gaussian primitives of the reconstruction in ambiguous areas are extracted and sampled for points to complete the point cloud. To address the challenges of large-scale reconstruction, we extend this pipeline with a divide-and-conquer scheme for building-sized point cloud completion. We evaluate on the task of LiDAR point cloud completion of synthetic and real-world scenes and find that our method outperforms previous reconstruction methods.

Paperid: 310, https://arxiv.org/pdf/2512.02972.pdf GitHub

Abstract:
Integrating LiDAR and camera information in the bird's eye view (BEV) representation has demonstrated its effectiveness in 3D object detection. However, because of the fundamental disparity in geometric accuracy between these sensors, indiscriminate fusion in previous methods often leads to degraded performance. In this paper, we propose BEVDilation, a novel LiDAR-centric framework that prioritizes LiDAR information in the fusion. By formulating image BEV features as implicit guidance rather than naive concatenation, our strategy effectively alleviates the spatial misalignment caused by image depth estimation errors. Furthermore, the image guidance can effectively help the LiDAR-centric paradigm to address the sparsity and semantic limitations of point clouds. Specifically, we propose a Sparse Voxel Dilation Block that mitigates the inherent point sparsity by densifying foreground voxels through image priors. Moreover, we introduce a Semantic-Guided BEV Dilation Block to enhance the LiDAR feature diffusion processing with image semantic guidance and long-range context capture. On the challenging nuScenes benchmark, BEVDilation achieves better performance than state-of-the-art methods while maintaining competitive computational efficiency. Importantly, our LiDAR-centric strategy demonstrates greater robustness to depth noise compared to naive fusion. The source code is available at https://github.com/gwenzhang/BEVDilation.

Paperid: 311, https://arxiv.org/pdf/2512.02965.pdf GitHub

Abstract:
In low-light environments like nighttime driving, image degradation severely challenges in-vehicle camera safety. Since existing enhancement algorithms are often too computationally intensive for vehicular applications, we propose UltraFast-LieNET, a lightweight multi-scale shifted convolutional network for real-time low-light image enhancement. We introduce a Dynamic Shifted Convolution (DSConv) kernel with only 12 learnable parameters for efficient feature extraction. By integrating DSConv with varying shift distances, a Multi-scale Shifted Residual Block (MSRB) is constructed to significantly expand the receptive field. To mitigate lightweight network instability, a residual structure and a novel multi-level gradient-aware loss function are incorporated. UltraFast-LieNET allows flexible parameter configuration, with a minimum size of only 36 parameters. Results on the LOLI-Street dataset show a PSNR of 26.51 dB, outperforming state-of-the-art methods by 4.6 dB while utilizing only 180 parameters. Experiments across four benchmark datasets validate its superior balance of real-time performance and enhancement quality under limited resources. Code is available at https://githubhttps://github.com/YuhanChen2024/UltraFast-LiNET

Paperid: 312, https://arxiv.org/pdf/2512.02947.pdf GitHub

Abstract:
While machine learning has enabled the rapid prediction of inorganic materials with novel properties, the challenge of determining how to synthesize these materials remains largely unsolved. Previous work has largely focused on predicting precursors or reaction conditions, but only rarely on full synthesis pathways. We introduce the ActionGraph, a directed acyclic graph framework that encodes both the chemical and procedural structure, in terms of synthesis operations, of inorganic synthesis reactions. Using 13,017 text-mined solid-state synthesis reactions from the Materials Project, we show that incorporating PCA-reduced ActionGraph adjacency matrices into a $k$-nearest neighbors retrieval model significantly improves synthesis pathway prediction. While the ActionGraph framework only results in a 1.34% and 2.76% increase in precursor and operation F1 scores (average over varying numbers of PCA components) respectively, the operation length matching accuracy rises 3.4 times (from 15.8% to 53.3%). We observe an interesting trade-off where precursor prediction performance peaks at 10-11 PCA components while operation prediction continues improving up to 30 components. This suggests composition information dominates precursor selection while structural information is critical for operation sequencing. Overall, the ActionGraph framework demonstrates strong potential, and with further adoption, its full range of benefits can be effectively realized.

Paperid: 313, https://arxiv.org/pdf/2512.02942.pdf GitHub GitHub

Abstract:
The next frontier for video generation lies in developing models capable of zero-shot reasoning, where understanding real-world scientific laws is crucial for accurate physical outcome modeling under diverse conditions. However, existing video benchmarks are physical commonsense-based, offering limited insight into video models' scientific reasoning capability. We introduce VideoScience-Bench, a benchmark designed to evaluate undergraduate-level scientific understanding in video models. Each prompt encodes a composite scientific scenario that requires understanding and reasoning across multiple scientific concepts to generate the correct phenomenon. The benchmark comprises 200 carefully curated prompts spanning 14 topics and 103 concepts in physics and chemistry. We conduct expert-annotated evaluations across seven state-of-the-art video models in T2V and I2V settings along five dimensions: Prompt Consistency, Phenomenon Congruency, Correct Dynamism, Immutability, and Spatio-Temporal Continuity. Using a VLM-as-a-Judge to assess video generations, we observe strong correlation with human assessments. To the best of our knowledge, VideoScience-Bench is the first benchmark to evaluate video models not only as generators but also as reasoners, requiring their generations to demonstrate scientific understanding consistent with expected physical and chemical phenomena. Our data and evaluation code are available at: \href{https://github.com/hao-ai-lab/VideoScience}{github.com/hao-ai-lab/VideoScience}.

Paperid: 314, https://arxiv.org/pdf/2512.02933.pdf GitHub

Abstract:
Text-guided video editing, particularly for object removal and addition, remains a challenging task due to the need for precise spatial and temporal consistency. Existing methods often rely on auxiliary masks or reference images for editing guidance, which limits their scalability and generalization. To address these issues, we propose LoVoRA, a novel framework for mask-free video object removal and addition using object-aware localization mechanism. Our approach utilizes a unique dataset construction pipeline that integrates image-to-video translation, optical flow-based mask propagation, and video inpainting, enabling temporally consistent edits. The core innovation of LoVoRA is its learnable object-aware localization mechanism, which provides dense spatio-temporal supervision for both object insertion and removal tasks. By leveraging a Diffusion Mask Predictor, LoVoRA achieves end-to-end video editing without requiring external control signals during inference. Extensive experiments and human evaluation demonstrate the effectiveness and high-quality performance of LoVoRA. https://cz-5f.github.io/LoVoRA.github.io

Paperid: 315, https://arxiv.org/pdf/2512.02873.pdf GitHub

Abstract:
Extensions of ω-automata to infinite alphabets typically rely on symbolic guards to keep the transition relation finite, and on registers or memory cells to preserve information from past symbols. Symbolic transitions alone are ill-suited to act on this information, and register automata have intricate formal semantics and issues with tractability. We propose a slightly different approach based on obligations, i.e., assignment-like constructs attached to transitions. Whenever a transition with an obligation is taken, the obligation is evaluated against the current symbol and yields a constraint on the next symbol that the automaton will read. We formalize obligation automata with existential and universal branching and Emerson-Lei acceptance conditions, which subsume classic families such as Büchi, Rabin, Strett, and parity automata. We show that these automata recognise a strict superset of ω-regular languages. To illustrate the practicality of our proposal, we also introduce a machine-readable format to express obligation automata and describe a tool implementing several operations over them, including automata product and emptiness checking.

Paperid: 316, https://arxiv.org/pdf/2512.02867.pdf GitHub

Abstract:
Cone-Beam Computed Tomography (CBCT) and Intraoral Scanning (IOS) are essential for digital dentistry, but annotated data scarcity limits automated solutions for pulp canal segmentation and cross-modal registration. To benchmark semi-supervised learning (SSL) in this domain, we organized the STSR 2025 Challenge at MICCAI 2025, featuring two tasks: (1) semi-supervised segmentation of teeth and pulp canals in CBCT, and (2) semi-supervised rigid registration of CBCT and IOS. We provided 60 labeled and 640 unlabeled IOS samples, plus 30 labeled and 250 unlabeled CBCT scans with varying resolutions and fields of view. The challenge attracted strong community participation, with top teams submitting open-source deep learning-based SSL solutions. For segmentation, leading methods used nnU-Net and Mamba-like State Space Models with pseudo-labeling and consistency regularization, achieving a Dice score of 0.967 and Instance Affinity of 0.738 on the hidden test set. For registration, effective approaches combined PointNetLK with differentiable SVD and geometric augmentation to handle modality gaps; hybrid neural-classical refinement enabled accurate alignment despite limited labels. All data and code are publicly available at https://github.com/ricoleehduu/STS-Challenge-2025 to ensure reproducibility.

Paperid: 317, https://arxiv.org/pdf/2512.02835.pdf GitHub

Abstract:
Reasoning-centric video object segmentation is an inherently complex task: the query often refers to dynamics, causality, and temporal interactions, rather than static appearances. Yet existing solutions generally collapse these factors into simplified reasoning with latent embeddings, rendering the reasoning chain opaque and essentially intractable. We therefore adopt an explicit decomposition perspective and introduce ReVSeg, which executes reasoning as sequential decisions in the native interface of pretrained vision language models (VLMs). Rather than folding all reasoning into a single-step prediction, ReVSeg executes three explicit operations -- semantics interpretation, temporal evidence selection, and spatial grounding -- aligning pretrained capabilities. We further employ reinforcement learning to optimize the multi-step reasoning chain, enabling the model to self-refine its decision quality from outcome-driven signals. Experimental results demonstrate that ReVSeg attains state-of-the-art performances on standard video object segmentation benchmarks and yields interpretable reasoning trajectories. Project page is available at https://clementine24.github.io/ReVSeg/ .

Paperid: 318, https://arxiv.org/pdf/2512.02817.pdf GitHub

Abstract:
The globalization of education and rapid growth of online learning have made localizing educational content a critical challenge. Lecture materials are inherently multimodal, combining spoken audio with visual slides, which requires systems capable of processing multiple input modalities. To provide an accessible and complete learning experience, translations must preserve all modalities: text for reading, slides for visual understanding, and speech for auditory learning. We present \textbf{BOOM}, a multimodal multilingual lecture companion that jointly translates lecture audio and slides to produce synchronized outputs across three modalities: translated text, localized slides with preserved visual elements, and synthesized speech. This end-to-end approach enables students to access lectures in their native language while aiming to preserve the original content in its entirety. Our experiments demonstrate that slide-aware transcripts also yield cascading benefits for downstream tasks such as summarization and question answering. We release our Slide Translation code at https://github.com/saikoneru/image-translator and integrate it in Lecture Translator at https://gitlab.kit.edu/kit/isl-ai4lt/lt-middleware/ltpipeline}\footnote{All released code and models are licensed under the MIT License.

Paperid: 319, https://arxiv.org/pdf/2512.02795.pdf GitHub

Abstract:
Code-generating LLMs are trained largely on static artifacts (source, comments, specifications) and rarely on materializations of run-time behavior. As a result, they readily internalize buggy or mislabeled code. Since non-trivial semantic properties are undecidable in general, the only practical way to obtain ground-truth functionality is by dynamic observation of executions. In prior work, we addressed representation with Sequence Sheets, Stimulus-Response Matrices (SRMs), and Stimulus-Response Cubes (SRCs) to capture and compare behavior across tests, implementations, and contexts. These structures make observation data analyzable offline and reusable, but they do not by themselves provide persistence, evolution, or interactive analytics at scale. In this paper, therefore, we introduce observation lakehouses that operationalize continual SRCs: a tall, append-only observations table storing every actuation (stimulus, response, context) and SQL queries that materialize SRC slices on demand. Built on Apache Parquet + Iceberg + DuckDB, the lakehouse ingests data from controlled pipelines (LASSO) and CI pipelines (e.g., unit test executions), enabling n-version assessment, behavioral clustering, and consensus oracles without re-execution. On a 509-problem benchmark, we ingest $\approx$8.6M observation rows ($<$51MiB) and reconstruct SRM/SRC views and clusters in $<$100ms on a laptop, demonstrating that continual behavior mining is practical without a distributed cluster of machines. This makes behavioral ground truth first-class alongside other run-time data and provides an infrastructure path toward behavior-aware evaluation and training. The Observation Lakehouse, together with the accompanying dataset, is publicly available as an open-source project on GitHub: https://github.com/SoftwareObservatorium/observation-lakehouse

Paperid: 320, https://arxiv.org/pdf/2512.02794.pdf GitHub

Abstract:
Recent diffusion-based text-to-image customization methods have achieved significant success in understanding concrete concepts to control generation processes, such as styles and shapes. However, few efforts dive into the realistic yet challenging customization of physical concepts. The core limitation of current methods arises from the absence of explicitly introducing physical knowledge during training. Even when physics-related words appear in the input text prompts, our experiments consistently demonstrate that these methods fail to accurately reflect the corresponding physical properties in the generated results. In this paper, we propose PhyCustom, a fine-tuning framework comprising two novel regularization losses to activate diffusion model to perform physical customization. Specifically, the proposed isometric loss aims at activating diffusion models to learn physical concepts while decouple loss helps to eliminate the mixture learning of independent concepts. Experiments are conducted on a diverse dataset and our benchmark results demonstrate that PhyCustom outperforms previous state-of-the-art and popular methods in terms of physical customization quantitatively and qualitatively.

Paperid: 321, https://arxiv.org/pdf/2512.02793.pdf GitHub

Abstract:
Video-based world models have recently garnered increasing attention for their ability to synthesize diverse and dynamic visual environments. In this paper, we focus on shared world modeling, where a model generates multiple videos from a set of input images, each representing the same underlying world in different camera poses. We propose IC-World, a novel generation framework, enabling parallel generation for all input images via activating the inherent in-context generation capability of large video models. We further finetune IC-World via reinforcement learning, Group Relative Policy Optimization, together with two proposed novel reward models to enforce scene-level geometry consistency and object-level motion consistency among the set of generated videos. Extensive experiments demonstrate that IC-World substantially outperforms state-of-the-art methods in both geometry and motion consistency. To the best of our knowledge, this is the first work to systematically explore the shared world modeling problem with video-based world models.

Paperid: 322, https://arxiv.org/pdf/2512.02792.pdf GitHub

Abstract:
Composed Video Retrieval (CVR) is a challenging video retrieval task that utilizes multi-modal queries, consisting of a reference video and modification text, to retrieve the desired target video. The core of this task lies in understanding the multi-modal composed query and achieving accurate composed feature learning. Within multi-modal queries, the video modality typically carries richer semantic content compared to the textual modality. However, previous works have largely overlooked the disparity in information density between these two modalities. This limitation can lead to two critical issues: 1) modification subject referring ambiguity and 2) limited detailed semantic focus, both of which degrade the performance of CVR models. To address the aforementioned issues, we propose a novel CVR framework, namely the Hierarchical Uncertainty-aware Disambiguation network (HUD). HUD is the first framework that leverages the disparity in information density between video and text to enhance multi-modal query understanding. It comprises three key components: (a) Holistic Pronoun Disambiguation, (b) Atomistic Uncertainty Modeling, and (c) Holistic-to-Atomistic Alignment. By exploiting overlapping semantics through holistic cross-modal interaction and fine-grained semantic alignment via atomistic-level cross-modal interaction, HUD enables effective object disambiguation and enhances the focus on detailed semantics, thereby achieving precise composed feature learning. Moreover, our proposed HUD is also applicable to the Composed Image Retrieval (CIR) task and achieves state-of-the-art performance across three benchmark datasets for both CVR and CIR tasks. The codes are available on https://zivchen-ty.github.io/HUD.github.io/.

Paperid: 323, https://arxiv.org/pdf/2512.02787.pdf GitHub

Abstract:
Vision-Language-Action (VLA) models have recently achieved remarkable progress in robotic manipulation, yet they remain limited in failure diagnosis and learning from failures. Additionally, existing failure datasets are mostly generated programmatically in simulation, which limits their generalization to the real world. In light of these, we introduce ViFailback, a framework designed to diagnose robotic manipulation failures and provide both textual and visual correction guidance. Our framework utilizes explicit visual symbols to enhance annotation efficiency. We further release the ViFailback dataset, a large-scale collection of 58,126 Visual Question Answering (VQA) pairs along with their corresponding 5,202 real-world manipulation trajectories. Based on the dataset, we establish ViFailback-Bench, a benchmark of 11 fine-grained VQA tasks designed to assess the failure diagnosis and correction abilities of Vision-Language Models (VLMs), featuring ViFailback-Bench Lite for closed-ended and ViFailback-Bench Hard for open-ended evaluation. To demonstrate the effectiveness of our framework, we built the ViFailback-8B VLM, which not only achieves significant overall performance improvement on ViFailback-Bench but also generates visual symbols for corrective action guidance. Finally, by integrating ViFailback-8B with a VLA model, we conduct real-world robotic experiments demonstrating its ability to assist the VLA model in recovering from failures. Project Website: https://x1nyuzhou.github.io/vifailback.github.io/

Paperid: 324, https://arxiv.org/pdf/2512.02786.pdf GitHub GitHub GitHub

Abstract:
Membership Inference Attacks (MIAs) aim to determine whether a specific data point was included in the training set of a target model. Although there are have been numerous methods developed for detecting data contamination in large language models (LLMs), their performance on multimodal LLMs (MLLMs) falls short due to the instabilities introduced through multimodal component adaptation and possible distribution shifts across multiple inputs. In this work, we investigate multimodal membership inference and address two issues: first, by identifying distribution shifts in the existing datasets, and second, by releasing an extended baseline pipeline to detect them. We also generalize the perturbation-based membership inference methods to MLLMs and release \textbf{FiMMIA} -- a modular \textbf{F}ramework for \textbf{M}ultimodal \textbf{MIA}.\footnote{The source code and framework have been made publicly available under the MIT license via \href{https://github.com/ai-forever/data_leakage_detect}{link}.The video demonstration is available on \href{https://youtu.be/a9L4-H80aSg}{YouTube}.} Our approach trains a neural network to analyze the target model's behavior on perturbed inputs, capturing distributional differences between members and non-members. Comprehensive evaluations on various fine-tuned multimodal models demonstrate the effectiveness of our perturbation-based membership inference attacks in multimodal domains.

Paperid: 325, https://arxiv.org/pdf/2512.02781.pdf GitHub

Abstract:
We present LumiX, a structured diffusion framework for coherent text-to-intrinsic generation. Conditioned on text prompts, LumiX jointly generates a comprehensive set of intrinsic maps (e.g., albedo, irradiance, normal, depth, and final color), providing a structured and physically consistent description of an underlying scene. This is enabled by two key contributions: 1) Query-Broadcast Attention, a mechanism that ensures structural consistency by sharing queries across all maps in each self-attention block. 2) Tensor LoRA, a tensor-based adaptation that parameter-efficiently models cross-map relations for efficient joint training. Together, these designs enable stable joint diffusion training and unified generation of multiple intrinsic properties. Experiments show that LumiX produces coherent and physically meaningful results, achieving 23% higher alignment and a better preference score (0.19 vs. -0.41) compared to the state of the art, and it can also perform image-conditioned intrinsic decomposition within the same framework.

Paperid: 326, https://arxiv.org/pdf/2512.02772.pdf GitHub

Abstract:
Large Language Models (LLMs) frequently exhibit hallucinations, generating content that appears fluent and coherent but is factually incorrect. Such errors undermine trust and hinder their adoption in real-world applications. To address this challenge, two distinct research paradigms have emerged: model-centric Hallucination Detection (HD) and text-centric Fact Verification (FV). Despite sharing the same goal, these paradigms have evolved in isolation, using distinct assumptions, datasets, and evaluation protocols. This separation has created a research schism that hinders their collective progress. In this work, we take a decisive step toward bridging this divide. We introduce UniFact, a unified evaluation framework that enables direct, instance-level comparison between FV and HD by dynamically generating model outputs and corresponding factuality labels. Through large-scale experiments across multiple LLM families and detection methods, we reveal three key findings: (1) No paradigm is universally superior; (2) HD and FV capture complementary facets of factual errors; and (3) hybrid approaches that integrate both methods consistently achieve state-of-the-art performance. Beyond benchmarking, we provide the first in-depth analysis of why FV and HD diverged, as well as empirical evidence supporting the need for their unification. The comprehensive experimental results call for a new, integrated research agenda toward unifying Hallucination Detection and Fact Verification in LLMs. We have open-sourced all the code, data, and baseline implementation at: https://github.com/oneal2000/UniFact/

Paperid: 327, https://arxiv.org/pdf/2512.02764.pdf GitHub

Abstract:
Parameter-Efficient Fine-Tuning (PEFT) methods address the increasing size of Large Language Models (LLMs). Currently, many newly introduced PEFT methods are challenging to replicate, deploy, or compare with one another. To address this, we introduce PEFT-Factory, a unified framework for efficient fine-tuning LLMs using both off-the-shelf and custom PEFT methods. While its modular design supports extensibility, it natively provides a representative set of 19 PEFT methods, 27 classification and text generation datasets addressing 12 tasks, and both standard and PEFT-specific evaluation metrics. As a result, PEFT-Factory provides a ready-to-use, controlled, and stable environment, improving replicability and benchmarking of PEFT methods. PEFT-Factory is a downstream framework that originates from the popular LLaMA-Factory, and is publicly available at https://github.com/kinit-sk/PEFT-Factory

Paperid: 328, https://arxiv.org/pdf/2512.02747.pdf GitHub

Abstract:
We present a simple $q$-ary family of single-error-correcting, double-error-detecting (SEC--DED) linear codes whose parity checks are tied directly to the base-$p$ ($q=p$ prime) digits of the coordinate index. For blocklength $n=p^r$ the construction uses only $r+1$ parity checks -- \emph{near-Hamming} overhead -- and admits an index-based decoder that runs in a single pass with constant-time location and magnitude recovery from the syndromes. Based on the prototype, we develop two extensions: Code A1, which removes specific redundant trits to achieve higher information rate and support variable-length encoding; and Code A2, which incorporates two group-sum checks together with a 3-wise XOR linear independence condition on index subsets, yielding a ternary distance-4 (SEC--TED) variant. Furthermore, we demonstrate how the framework generalizes via $n$-wise XOR linearly independent sets to construct codes with distance $d = n + 1$, notably recovering the ternary Golay code for $n = 5$ -- showing both structural generality and a serendipitous link to optimal classical codes. Our contribution is not optimality but \emph{implementational simplicity} and an \emph{array-friendly} structure: the checks are digitwise and global sums, the mapping from syndromes to error location is explicit, and the SEC--TED upgrade is modular. We position the scheme against classical $q$-ary Hamming and SPC/product-code baselines and provide a small comparison of parity overhead, decoding work, and two-error behavior.

Paperid: 329, https://arxiv.org/pdf/2512.02727.pdf GitHub

Abstract:
Modeling daily hand interactions often struggles with severe occlusions, such as when two hands overlap, which highlights the need for robust feature learning in 3D hand pose estimation (HPE). To handle such occluded hand images, it is vital to effectively learn the relationship between local image features (e.g., for occluded joints) and global context (e.g., cues from inter-joints, inter-hands, or the scene). However, most current 3D HPE methods still rely on ResNet for feature extraction, and such CNN's inductive bias may not be optimal for 3D HPE due to its limited capability to model the global context. To address this limitation, we propose an effective and efficient framework for visual feature extraction in 3D HPE using recent state space modeling (i.e., Mamba), dubbed Deformable Mamba (DF-Mamba). DF-Mamba is designed to capture global context cues beyond standard convolution through Mamba's selective state modeling and the proposed deformable state scanning. Specifically, for local features after convolution, our deformable scanning aggregates these features within an image while selectively preserving useful cues that represent the global context. This approach significantly improves the accuracy of structured 3D HPE, with comparable inference speed to ResNet-50. Our experiments involve extensive evaluations on five divergent datasets including single-hand and two-hand scenarios, hand-only and hand-object interactions, as well as RGB and depth-based estimation. DF-Mamba outperforms the latest image backbones, including VMamba and Spatial-Mamba, on all datasets and achieves state-of-the-art performance.

Paperid: 330, https://arxiv.org/pdf/2512.02716.pdf GitHub

Abstract:
Mental health conditions affect hundreds of millions globally, yet early detection remains limited. While large language models (LLMs) have shown promise in mental health applications, their size and computational demands hinder practical deployment. Small language models (SLMs) offer a lightweight alternative, but their use for social media--based mental health prediction remains largely underexplored. In this study, we introduce Menta, the first optimized SLM fine-tuned specifically for multi-task mental health prediction from social media data. Menta is jointly trained across six classification tasks using a LoRA-based framework, a cross-dataset strategy, and a balanced accuracy--oriented loss. Evaluated against nine state-of-the-art SLM baselines, Menta achieves an average improvement of 15.2\% across tasks covering depression, stress, and suicidality compared with the best-performing non--fine-tuned SLMs. It also achieves higher accuracy on depression and stress classification tasks compared to 13B-parameter LLMs, while being approximately 3.25x smaller. Moreover, we demonstrate real-time, on-device deployment of Menta on an iPhone 15 Pro Max, requiring only approximately 3GB RAM. Supported by a comprehensive benchmark against existing SLMs and LLMs, Menta highlights the potential for scalable, privacy-preserving mental health monitoring. Code is available at: https://xxue752-nz.github.io/menta-project/

Paperid: 331, https://arxiv.org/pdf/2512.02697.pdf GitHub

Abstract:
Cross-view geo-localization infers a location by retrieving geo-tagged reference images that visually correspond to a query image. However, the traditional satellite-centric paradigm limits robustness when high-resolution or up-to-date satellite imagery is unavailable. It further underexploits complementary cues across views (e.g., drone, satellite, and street) and modalities (e.g., language and image). To address these challenges, we propose GeoBridge, a foundation model that performs bidirectional matching across views and supports language-to-image retrieval. Going beyond traditional satellite-centric formulations, GeoBridge builds on a novel semantic-anchor mechanism that bridges multi-view features through textual descriptions for robust, flexible localization. In support of this task, we construct GeoLoc, the first large-scale, cross-modal, and multi-view aligned dataset comprising over 50,000 pairs of drone, street-view panorama, and satellite images as well as their textual descriptions, collected from 36 countries, ensuring both geographic and semantic alignment. We performed broad evaluations across multiple tasks. Experiments confirm that GeoLoc pre-training markedly improves geo-location accuracy for GeoBridge while promoting cross-domain generalization and cross-modal knowledge transfer. The dataset, source code, and pretrained models were released at https://github.com/MiliLab/GeoBridge.

Paperid: 332, https://arxiv.org/pdf/2512.02681.pdf GitHub

Abstract:
Although diffusion-based models have achieved impressive results in image super-resolution, they often rely on large-scale backbones such as Stable Diffusion XL (SDXL) and Diffusion Transformers (DiT), which lead to excessive computational and memory costs during training and inference. To address this issue, we develop a lightweight diffusion method, PGP-DiffSR, by removing redundant information from diffusion models under the guidance of the phase information of inputs for efficient image super-resolution. We first identify the intra-block redundancy within the diffusion backbone and propose a progressive pruning approach that removes redundant blocks while reserving restoration capability. We note that the phase information of the restored images produced by the pruned diffusion model is not well estimated. To solve this problem, we propose a phase-exchange adapter module that explores the phase information of the inputs to guide the pruned diffusion model for better restoration performance. We formulate the progressive pruning approach and the phase-exchange adapter module into a unified model. Extensive experiments demonstrate that our method achieves competitive restoration quality while significantly reducing computational load and memory consumption. The code is available at https://github.com/yzb1997/PGP-DiffSR.

Paperid: 333, https://arxiv.org/pdf/2512.02660.pdf GitHub

Abstract:
Vision-language models (VLMs) like ColPali achieve state-of-the-art document retrieval by embedding pages as images and computing fine-grained similarity between query tokens and visual patches. However, they return entire pages rather than specific regions, limiting utility for retrieval-augmented generation (RAG) where precise context is paramount. Conversely, OCR-based systems extract structured text with bounding box coordinates but lack semantic grounding for relevance assessment. We propose a hybrid architecture that unifies these paradigms: using ColPali's patch-level similarity scores as spatial relevance filters over OCR-extracted regions. We formalize the coordinate mapping between vision transformer patch grids and OCR bounding boxes, introduce intersection metrics for relevance propagation, and establish theoretical bounds on retrieval precision. Our approach operates at inference time without additional training. We release Snappy, an open-source implementation demonstrating practical applicability, with empirical evaluation ongoing.

Paperid: 334, https://arxiv.org/pdf/2512.02650.pdf GitHub

Abstract:
This work introduces a new task, text-conditioned selective video-to-audio (V2A) generation, which produces only the user-intended sound from a multi-object video. This capability is especially crucial in multimedia production, where audio tracks are handled individually for each sound source for precise editing, mixing, and creative control. However, current approaches generate single source-mixed sounds at once, largely because visual features are entangled, and region cues or prompts often fail to specify the source. We propose SelVA, a novel text-conditioned V2A model that treats the text prompt as an explicit selector of target source and modulates video encoder to distinctly extract prompt-relevant video features. The proposed supplementary tokens promote cross-attention by suppressing text-irrelevant activations with efficient parameter tuning, yielding robust semantic and temporal grounding. SelVA further employs a self-augmentation scheme to overcome the lack of mono audio track supervision. We evaluate SelVA on VGG-MONOAUDIO, a curated benchmark of clean single-source videos for such a task. Extensive experiments and ablations consistently verify its effectiveness across audio quality, semantic alignment, and temporal synchronization. Code and demo are available at https://jnwnlee.github.io/selva-demo/.

Paperid: 335, https://arxiv.org/pdf/2512.02648.pdf GitHub

Abstract:
We introduce PoreTrack3D, the first benchmark for dynamic 3D Gaussian splatting in pore-scale, non-rigid 3D facial trajectory tracking. It contains over 440,000 facial trajectories in total, among which more than 52,000 are longer than 10 frames, including 68 manually reviewed trajectories that span the entire 150 frames. To the best of our knowledge, PoreTrack3D is the first benchmark dataset to capture both traditional facial landmarks and pore-scale keypoints trajectory, advancing the study of fine-grained facial expressions through the analysis of subtle skin-surface motion. We systematically evaluate state-of-the-art dynamic 3D Gaussian splatting methods on PoreTrack3D, establishing the first performance baseline in this domain. Overall, the pipeline developed for this benchmark dataset's creation establishes a new framework for high-fidelity facial motion capture and dynamic 3D reconstruction. Our dataset are publicly available at: https://github.com/JHXion9/PoreTrack3D

Paperid: 336, https://arxiv.org/pdf/2512.02610.pdf GitHub

Abstract:
Accurate prediction of the Remaining Useful Life (RUL) in machinery can significantly diminish maintenance costs, enhance equipment up-time, and mitigate adverse outcomes. Data-driven RUL prediction techniques have demonstrated commendable performance. However, their efficacy often relies on the assumption that training and testing data are drawn from the same distribution or domain, which does not hold in real industrial settings. To mitigate this domain discrepancy issue, prior adversarial domain adaptation methods focused on deriving domain-invariant features. Nevertheless, they overlook target-specific information and inconsistency characteristics pertinent to the degradation stages, resulting in suboptimal performance. To tackle these issues, we propose a novel domain adaptation approach for cross-domain RUL prediction named TACDA. Specifically, we propose a target domain reconstruction strategy within the adversarial adaptation process, thereby retaining target-specific information while learning domain-invariant features. Furthermore, we develop a novel clustering and pairing strategy for consistent alignment between similar degradation stages. Through extensive experiments, our results demonstrate the remarkable performance of our proposed TACDA method, surpassing state-of-the-art approaches with regard to two different evaluation metrics. Our code is available at https://github.com/keyplay/TACDA.

Paperid: 337, https://arxiv.org/pdf/2512.02589.pdf GitHub

Abstract:
Large language models are increasingly embedded into academic writing workflows, yet existing assistants remain external to the editor, preventing deep interaction with document state, structure, and revision history. This separation makes it impossible to support agentic, context-aware operations directly within LaTeX editors such as Overleaf. We present PaperDebugger, an in-editor, multi-agent, and plugin-based academic writing assistant that brings LLM-driven reasoning directly into the writing environment. Enabling such in-editor interaction is technically non-trivial: it requires reliable bidirectional synchronization with the editor, fine-grained version control and patching, secure state management, multi-agent scheduling, and extensible communication with external tools. PaperDebugger addresses these challenges through a Chrome-approved extension, a Kubernetes-native orchestration layer, and a Model Context Protocol (MCP) toolchain that integrates literature search, reference lookup, document scoring, and revision pipelines. Our demo showcases a fully integrated workflow, including localized edits, structured reviews, parallel agent execution, and diff-based updates, encapsulated within a minimal-intrusion user interface (UI). Early aggregated analytics demonstrate active user engagement and validate the practicality of an editor-native, agentic writing assistant. More details about this demo and video could be found at https://github.com/PaperDebugger/PaperDebugger.

Paperid: 338, https://arxiv.org/pdf/2512.02547.pdf GitHub

Abstract:
Many approximations were suggested to circumvent the cubic complexity of kernel-based algorithms, allowing their application to large-scale datasets. One strategy is to consider the primal formulation of the learning problem by mapping the data to a higher-dimensional space using tensor-product structured polynomial and Fourier features. The curse of dimensionality due to these tensor-product features was effectively solved by a tensor network reparameterization of the model parameters. However, another important aspect of model training - identifying optimal feature hyperparameters - has not been addressed and is typically handled using the standard cross-validation approach. In this paper, we introduce the Feature Learning (FL) model, which addresses this issue by representing tensor-product features as a learnable Canonical Polyadic Decomposition (CPD). By leveraging this CPD structure, we efficiently learn the hyperparameters associated with different features alongside the model parameters using an Alternating Least Squares (ALS) optimization method. We prove the effectiveness of the FL model through experiments on real data of various dimensionality and scale. The results show that the FL model can be consistently trained 3-5 times faster than and have the prediction quality on par with a standard cross-validated model.

Paperid: 339, https://arxiv.org/pdf/2512.02535.pdf GitHub

Abstract:
Information gathering in large-scale or time-critical scenarios (e.g., environmental monitoring, search and rescue) requires broad coverage within limited time budgets, motivating the use of multi-agent systems. These scenarios are commonly formulated as multi-agent informative path planning (MAIPP), where multiple agents must coordinate to maximize information gain while operating under budget constraints. A central challenge in MAIPP is ensuring effective coordination while the belief over the environment evolves with incoming measurements. Recent learning-based approaches address this by using distributions over future positions as "intent" to support coordination. However, these autoregressive intent predictors are computationally expensive and prone to compounding errors. Inspired by the effectiveness of diffusion models as expressive, long-horizon policies, we propose AID, a fully decentralized MAIPP framework that leverages diffusion models to generate long-term trajectories in a non-autoregressive manner. AID first performs behavior cloning on trajectories produced by existing MAIPP planners and then fine-tunes the policy using reinforcement learning via Diffusion Policy Policy Optimization (DPPO). This two-stage pipeline enables the policy to inherit expert behavior while learning improved coordination through online reward feedback. Experiments demonstrate that AID consistently improves upon the MAIPP planners it is trained from, achieving up to 4x faster execution and 17% increased information gain, while scaling effectively to larger numbers of agents. Our implementation is publicly available at https://github.com/marmotlab/AID.

Paperid: 340, https://arxiv.org/pdf/2512.02532.pdf GitHub

Abstract:
To address the scalability limitations of Gaussian process (GP) regression, several approximation techniques have been proposed. One such method is based on tensor networks, which utilizes an exponential number of basis functions without incurring exponential computational cost. However, extending this model to a fully probabilistic formulation introduces several design challenges. In particular, for tensor train (TT) models, it is unclear which TT-core should be treated in a Bayesian manner. We introduce a Bayesian tensor train kernel machine that applies Laplace approximation to estimate the posterior distribution over a selected TT-core and employs variational inference (VI) for precision hyperparameters. Experiments show that core selection is largely independent of TT-ranks and feature structure, and that VI replaces cross-validation while offering up to 65x faster training. The method's effectiveness is demonstrated on an inverse dynamics problem.

Paperid: 341, https://arxiv.org/pdf/2512.02530.pdf GitHub

Abstract:
The exponential growth of digital content presents significant challenges for content safety. Current moderation systems, often based on single models or fixed pipelines, exhibit limitations in identifying implicit risks and providing interpretable judgment processes. To address these issues, we propose Aetheria, a multimodal interpretable content safety framework based on multi-agent debate and collaboration.Employing a collaborative architecture of five core agents, Aetheria conducts in-depth analysis and adjudication of multimodal content through a dynamic, mutually persuasive debate mechanism, which is grounded by RAG-based knowledge retrieval.Comprehensive experiments on our proposed benchmark (AIR-Bench) validate that Aetheria not only generates detailed and traceable audit reports but also demonstrates significant advantages over baselines in overall content safety accuracy, especially in the identification of implicit risks. This framework establishes a transparent and interpretable paradigm, significantly advancing the field of trustworthy AI content moderation.

Paperid: 342, https://arxiv.org/pdf/2512.02523.pdf GitHub

Abstract:
Singing voice synthesis (SVS) has advanced significantly, enabling models to generate vocals with accurate pitch and consistent style. As these capabilities improve, the need for reliable evaluation and optimization becomes increasingly critical. However, current methods like reward systems often rely on single numerical scores, struggle to capture various dimensions such as phrasing or expressiveness, and require costly annotations, limiting interpretability and generalization. To address these issues, we propose a generative feedback (i.e., reward model) framework that provides multi-dimensional language and audio feedback for SVS assessment. Our approach leverages an audio-language model to generate text and audio critiques-covering aspects such as melody, content, and auditory quality. The model is fine-tuned on a hybrid dataset combining human music reactions and synthetic critiques from a MLLMs, enhancing diversity and linguistic richness. Quantitative experiments validate the effectiveness of the proposed dataset and training strategy, demonstrating that the framework produces musically accurate and interpretable evaluations suitable for guiding generative model improvement. The code is at [https://github.com/opendilab/VocalCritic](https://github.com/opendilab/VocalCritic)

Paperid: 343, https://arxiv.org/pdf/2512.02497.pdf GitHub

Abstract:
Test time Adaptation is a promising approach for mitigating domain shift in medical image segmentation; however, current evaluations remain limited in terms of modality coverage, task diversity, and methodological consistency. We present MedSeg-TTA, a comprehensive benchmark that examines twenty representative adaptation methods across seven imaging modalities, including MRI, CT, ultrasound, pathology, dermoscopy, OCT, and chest X-ray, under fully unified data preprocessing, backbone configuration, and test time protocols. The benchmark encompasses four significant adaptation paradigms: Input-level Transformation, Feature-level Alignment, Output-level Regularization, and Prior Estimation, enabling the first systematic cross-modality comparison of their reliability and applicability. The results show that no single paradigm performs best in all conditions. Input-level methods are more stable under mild appearance shifts. Feature-level and Output-level methods offer greater advantages in boundary-related metrics, whereas prior-based methods exhibit strong modality dependence. Several methods degrade significantly under large inter-center and inter-device shifts, which highlights the importance of principled method selection for clinical deployment. MedSeg-TTA provides standardized datasets, validated implementations, and a public leaderboard, establishing a rigorous foundation for future research on robust, clinically reliable test-time adaptation. All source codes and open-source datasets are available at https://github.com/wenjing-gg/MedSeg-TTA.

Paperid: 344, https://arxiv.org/pdf/2512.02494.pdf GitHub

Abstract:
Differentiable optimization layers enable learning systems to make decisions by solving embedded optimization problems. However, computing gradients via implicit differentiation requires solving a linear system with Hessian terms, which is both compute- and memory-intensive. To address this challenge, we propose a novel algorithm that computes the gradient using only first-order information. The key insight is to rewrite the differentiable optimization as a bilevel optimization problem and leverage recent advances in bilevel methods. Specifically, we introduce an active-set Lagrangian hypergradient oracle that avoids Hessian evaluations and provides finite-time, non-asymptotic approximation guarantees. We show that an approximate hypergradient can be computed using only first-order information in $\tilde{\oo}(1)$ time, leading to an overall complexity of $\tilde{\oo}(δ^{-1}ε^{-3})$ for constrained bilevel optimization, which matches the best known rate for non-smooth non-convex optimization. Furthermore, we release an open-source Python library that can be easily adapted from existing solvers. Our code is available here: https://github.com/guaguakai/FFOLayer.

Paperid: 345, https://arxiv.org/pdf/2512.02492.pdf GitHub

Abstract:
While diffusion model for audio-driven avatar video generation have achieved notable process in synthesizing long sequences with natural audio-visual synchronization and identity consistency, the generation of music-performance videos with camera motions remains largely unexplored. We present YingVideo-MV, the first cascaded framework for music-driven long-video generation. Our approach integrates audio semantic analysis, an interpretable shot planning module (MV-Director), temporal-aware diffusion Transformer architectures, and long-sequence consistency modeling to enable automatic synthesis of high-quality music performance videos from audio signals. We construct a large-scale Music-in-the-Wild Dataset by collecting web data to support the achievement of diverse, high-quality results. Observing that existing long-video generation methods lack explicit camera motion control, we introduce a camera adapter module that embeds camera poses into latent noise. To enhance continulity between clips during long-sequence inference, we further propose a time-aware dynamic window range strategy that adaptively adjust denoising ranges based on audio embedding. Comprehensive benchmark tests demonstrate that YingVideo-MV achieves outstanding performance in generating coherent and expressive music videos, and enables precise music-motion-camera synchronization. More videos are available in our project page: https://giantailab.github.io/YingVideo-MV/ .

Paperid: 346, https://arxiv.org/pdf/2512.02485.pdf GitHub

Abstract:
Vision-Language Models (VLMs) show promise in medical diagnosis, yet suffer from reasoning detachment, where linguistically fluent explanations drift from verifiable image evidence, undermining clinical trust. Recent multi-agent frameworks simulate Multidisciplinary Team (MDT) debates to mitigate single-model bias, but open-ended discussions amplify textual noise and computational cost while failing to anchor reasoning to visual evidence, the cornerstone of medical decision-making. We propose UCAgents, a hierarchical multi-agent framework enforcing unidirectional convergence through structured evidence auditing. Inspired by clinical workflows, UCAgents forbids position changes and limits agent interactions to targeted evidence verification, suppressing rhetorical drift while amplifying visual signal extraction. In UCAgents, a one-round inquiry discussion is introduced to uncover potential risks of visual-textual misalignment. This design jointly constrains visual ambiguity and textual noise, a dual-noise bottleneck that we formalize via information theory. Extensive experiments on four medical VQA benchmarks show UCAgents achieves superior accuracy (71.3% on PathVQA, +6.0% over state-of-the-art) with 87.7% lower token cost, the evaluation results further confirm that UCAgents strikes a balance between uncovering more visual evidence and avoiding confusing textual interference. These results demonstrate that UCAgents exhibits both diagnostic reliability and computational efficiency critical for real-world clinical deployment. Code is available at https://github.com/fqhank/UCAgents.

Paperid: 347, https://arxiv.org/pdf/2512.02457.pdf GitHub

Abstract:
Recent audio-video generative systems suggest that coupling modalities benefits not only audio-video synchrony but also the video modality itself. We pose a fundamental question: Does audio-video joint denoising training improve video generation, even when we only care about video quality? To study this, we introduce a parameter-efficient Audio-Video Full DiT (AVFullDiT) architecture that leverages pre-trained text-to-video (T2V) and text-to-audio (T2A) modules for joint denoising. We train (i) a T2AV model with AVFullDiT and (ii) a T2V-only counterpart under identical settings. Our results provide the first systematic evidence that audio-video joint denoising can deliver more than synchrony. We observe consistent improvements on challenging subsets featuring large and object contact motions. We hypothesize that predicting audio acts as a privileged signal, encouraging the model to internalize causal relationships between visual events and their acoustic consequences (e.g., collision $\times$ impact sound), which in turn regularizes video dynamics. Our findings suggest that cross-modal co-training is a promising approach to developing stronger, more physically grounded world models. Code and dataset will be made publicly available.

Paperid: 348, https://arxiv.org/pdf/2512.02438.pdf GitHub

Abstract:
In medical healthcare, obtaining detailed annotations is challenging, highlighting the need for robust Vision-Language Models (VLMs). Pretrained VLMs enable fine-tuning on small datasets or zero-shot inference, achieving performance comparable to task-specific models. Contrastive learning (CL) is a key paradigm for training VLMs but inherently requires large batch sizes for effective learning, making it computationally demanding and often limited to well-resourced institutions. Moreover, with limited data in healthcare, it is important to prioritize knowledge extraction from both data and models during training to improve performance. Therefore, we focus on leveraging the momentum method combined with distillation to simultaneously address computational efficiency and knowledge exploitation. Our contributions can be summarized as follows: (1) leveraging momentum self-distillation to enhance multimodal learning, and (2) integrating momentum mechanisms with gradient accumulation to enlarge the effective batch size without increasing resource consumption. Our method attains competitive performance with state-of-the-art (SOTA) approaches in zero-shot classification, while providing a substantial boost in the few-shot adaption, achieving over 90% AUC-ROC and improving retrieval tasks by 2-3%. Importantly, our method achieves high training efficiency with a single GPU while maintaining reasonable training time. Our approach aims to advance efficient multimodal learning by reducing resource requirements while improving performance over SOTA methods. The implementation of our method is available at https://github.com/phphuc612/MSD .

Paperid: 349, https://arxiv.org/pdf/2512.02400.pdf GitHub GitHub

Abstract:
Object-goal navigation in open-vocabulary settings requires agents to locate novel objects in unseen environments, yet existing approaches suffer from opaque decision-making processes and low success rate on locating unseen objects. To address these challenges, we propose Nav-$R^2$, a framework that explicitly models two critical types of relationships, target-environment modeling and environment-action planning, through structured Chain-of-Thought (CoT) reasoning coupled with a Similarity-Aware Memory. We construct a Nav$R^2$-CoT dataset that teaches the model to perceive the environment, focus on target-related objects in the surrounding context and finally make future action plans. Our SA-Mem preserves the most target-relevant and current observation-relevant features from both temporal and semantic perspectives by compressing video frames and fusing historical observations, while introducing no additional parameters. Compared to previous methods, Nav-R^2 achieves state-of-the-art performance in localizing unseen objects through a streamlined and efficient pipeline, avoiding overfitting to seen object categories while maintaining real-time inference at 2Hz. Resources will be made publicly available at \href{https://github.com/AMAP-EAI/Nav-R2}{github link}.

Paperid: 350, https://arxiv.org/pdf/2512.02392.pdf GitHub

Abstract:
End-to-end multi-object tracking (MOT) methods have recently achieved remarkable progress by unifying detection and association within a single framework. Despite their strong detection performance, these methods suffer from relatively low association accuracy. Through detailed analysis, we observe that object embeddings produced by the shared DETR architecture display excessively high inter-object similarity, as it emphasizes only category-level discrimination within single frames. In contrast, tracking requires instance-level distinction across frames with spatial and temporal continuity, for which current end-to-end approaches insufficiently optimize object embeddings. To address this, we introduce FDTA (From Detection to Association), an explicit feature refinement framework that enhances object discriminativeness across three complementary perspectives. Specifically, we introduce a Spatial Adapter (SA) to integrate depth-aware cues for spatial continuity, a Temporal Adapter (TA) to aggregate historical information for temporal dependencies, and an Identity Adapter (IA) to leverage quality-aware contrastive learning for instance-level separability. Extensive experiments demonstrate that FDTA achieves state-of-the-art performance on multiple challenging MOT benchmarks, including DanceTrack, SportsMOT, and BFT, highlighting the effectiveness of our proposed discriminative embedding enhancement strategy. The code is available at https://github.com/Spongebobbbbbbbb/FDTA.

Paperid: 351, https://arxiv.org/pdf/2512.02375.pdf GitHub

Abstract:
Compared with conventional offline UAV photogrammetry, real-time UAV photogrammetry is essential for time-critical geospatial applications such as disaster response and active digital-twin maintenance. However, most existing methods focus on processing captured images or sequential frames in real time, without explicitly evaluating the quality of the on-the-go 3D reconstruction or providing guided feedback to enhance image acquisition in the target area. This work presents On-the-fly Feedback SfM, an explore-and-exploit framework for real-time UAV photogrammetry, enabling iterative exploration of unseen regions and exploitation of already observed and reconstructed areas in near real time. Built upon SfM on-the-fly , the proposed method integrates three modules: (1) online incremental coarse-mesh generation for dynamically expanding sparse 3D point cloud; (2) online mesh quality assessment with actionable indicators; and (3) predictive path planning for on-the-fly trajectory refinement. Comprehensive experiments demonstrate that our method achieves in-situ reconstruction and evaluation in near real time while providing actionable feedback that markedly reduces coverage gaps and re-flight costs. Via the integration of data collection, processing, 3D reconstruction and assessment, and online feedback, our on the-fly feedback SfM could be an alternative for the transition from traditional passive working mode to a more intelligent and adaptive exploration workflow. Code is now available at https://github.com/IRIS-LAB-whu/OntheflySfMFeedback.

Paperid: 352, https://arxiv.org/pdf/2512.02359.pdf GitHub

Abstract:
Multi-view crowd counting can effectively mitigate occlusion issues that commonly arise in single-image crowd counting. Existing deep-learning multi-view crowd counting methods project different camera view images onto a common space to obtain ground-plane density maps, requiring abundant and costly crowd annotations and camera calibrations. Hence, calibration-free methods are proposed that do not require camera calibrations and scene-level crowd annotations. However, existing calibration-free methods still require expensive image-level crowd annotations for training the single-view counting module. Thus, in this paper, we propose a weakly-supervised calibration-free multi-view crowd counting method (WSCF-MVCC), directly using crowd count as supervision for the single-view counting module rather than density maps constructed from crowd annotations. Instead, a self-supervised ranking loss that leverages multi-scale priors is utilized to enhance the model's perceptual ability without additional annotation costs. What's more, the proposed model leverages semantic information to achieve a more accurate view matching and, consequently, a more precise scene-level crowd count estimation. The proposed method outperforms the state-of-the-art methods on three widely used multi-view counting datasets under weakly supervised settings, indicating that it is more suitable for practical deployment compared with calibrated methods. Code is released in https://github.com/zqyq/Weakly-MVCC.

Paperid: 353, https://arxiv.org/pdf/2512.02351.pdf GitHub GitHub

Abstract:
Large multimodal models have achieved remarkable progress in both understanding and generation. Recent efforts pursue unified multimodal models that integrate heterogeneous components to support both capabilities within a single framework. However, such unification introduces inference inefficiencies, e.g., specific tasks or samples may not require the full knowledge or capacity of the unified model. Yet, a systematic understanding of how these inefficiencies manifest across different components remains limited. In this work, we first conduct a systematic analysis of unified multimodal model components using training-free pruning as a probing methodology, considering both depth pruning and width reduction. Our study reveals that the understanding component exhibits notable compressibility in both understanding and generation tasks, which is more pronounced in the latter. In contrast, the generation components are highly sensitive to compression, with performance deteriorating sharply even under moderate compression ratios. To address this limitation, we propose the Mixture-of-Experts (MoE) Adaptation, inspired by the dynamic activation patterns observed across different samples. This approach partitions the generation module into multiple experts and enables sparse activation to restore generation quality. We validate the effectiveness of sparse activation through expert-frozen tuning and further demonstrate that a fully trainable adaptation delivers additional gains. As a result, the adapted BAGEL model achieves performance comparable to the full model while activating only about half of its parameters. The code is released at \href{https://github.com/Shwai-He/SparseUnifiedModel}{this link}.

Paperid: 354, https://arxiv.org/pdf/2512.02341.pdf GitHub GitHub

Abstract:
3D vision foundation models have shown strong generalization in reconstructing key 3D attributes from uncalibrated images through a single feed-forward pass. However, when deployed in online settings such as driving scenarios, predictions are made over temporal windows, making it non-trivial to maintain consistency across time. Recent strategies align consecutive predictions by solving global transformation, yet our analysis reveals their fundamental limitations in assumption validity, local alignment scope, and robustness under noisy geometry. In this work, we propose a higher-DOF and long-term alignment framework based on Thin Plate Spline, leveraging globally propagated control points to correct spatially varying inconsistencies. In addition, we adopt a point-agnostic submap registration design that is inherently robust to noisy geometry predictions. The proposed framework is fully plug-and-play, compatible with diverse 3D foundation models and camera configurations (e.g., monocular or surround-view). Extensive experiments demonstrate that our method consistently yields more coherent geometry and lower trajectory errors across multiple datasets, backbone models, and camera setups, highlighting its robustness and generality. Codes are publicly available at \href{https://github.com/Xian-Bei/TALO}{https://github.com/Xian-Bei/TALO}.

Paperid: 355, https://arxiv.org/pdf/2512.02340.pdf GitHub

Abstract:
Spatial reasoning is a core aspect of human intelligence that allows perception, inference and planning in 3D environments. However, current vision-language models (VLMs) struggle to maintain geometric coherence and cross-view consistency for spatial reasoning in multi-view settings. We attribute this gap to the lack of fine-grained benchmarks that isolate multi-view reasoning from single-view perception and temporal factors. To address this, we present ReMindView-Bench, a cognitively grounded benchmark for evaluating how VLMs construct, align and maintain spatial mental models across complementary viewpoints. ReMindView-Bench systematically varies viewpoint spatial pattern and query type to probe key factors of spatial cognition. Evaluations of 15 current VLMs reveals consistent failures in cross-view alignment and perspective-taking in multi-view spatial reasoning, motivating deeper analysis on the reasoning process. Explicit phase-wise analysis using LLM-as-a-judge and self-consistency prompting shows that VLMs perform well on in-frame perception but degrade sharply when integrating information across views. Implicit analysis, including linear probing and entropy dynamics, further show progressive loss of task-relevant information and uncertainty separation between correct and incorrect trajectories. These results provide a cognitively grounded diagnosis of VLM spatial reasoning and reveal how multi-view spatial mental models are formed, degraded and destabilized across reasoning phases. The ReMindView-Bench benchmark is available at https://huggingface.co/datasets/Xue0823/ReMindView-Bench, and the source codes of benchmark construction and VLM reasoning analysis are available at https://github.com/pittisl/ReMindView-Bench.

Paperid: 356, https://arxiv.org/pdf/2512.02268.pdf GitHub

Abstract:
Generative models have the potential to transform the way we emulate Earth's changing climate. Previous generative approaches rely on weather-scale autoregression for climate emulation, but this is inherently slow for long climate horizons and has yet to demonstrate stable rollouts under nonstationary forcings. Here, we introduce Spatiotemporal Pyramid Flows (SPF), a new class of flow matching approaches that model data hierarchically across spatial and temporal scales. Inspired by cascaded video models, SPF partitions the generative trajectory into a spatiotemporal pyramid, progressively increasing spatial resolution to reduce computation and coupling each stage with an associated timescale to enable direct sampling at any temporal level in the pyramid. This design, together with conditioning each stage on prescribed physical forcings (e.g., greenhouse gases or aerosols), enables efficient, parallel climate emulation at multiple timescales. On ClimateBench, SPF outperforms strong flow matching baselines and pre-trained models at yearly and monthly timescales while offering fast sampling, especially at coarser temporal levels. To scale SPF, we curate ClimateSuite, the largest collection of Earth system simulations to date, comprising over 33,000 simulation-years across ten climate models and the first dataset to include simulations of climate interventions. We find that the scaled SPF model demonstrates good generalization to held-out scenarios across climate models. Together, SPF and ClimateSuite provide a foundation for accurate, efficient, probabilistic climate emulation across temporal scales and realistic future scenarios. Data and code is publicly available at https://github.com/stanfordmlgroup/spf .

Paperid: 357, https://arxiv.org/pdf/2512.02261.pdf GitHub

Abstract:
LLM-based trading agents are increasingly deployed in real-world financial markets to perform autonomous analysis and execution. However, their reliability and robustness under adversarial or faulty conditions remain largely unexamined, despite operating in high-risk, irreversible financial environments. We propose TradeTrap, a unified evaluation framework for systematically stress-testing both adaptive and procedural autonomous trading agents. TradeTrap targets four core components of autonomous trading agents: market intelligence, strategy formulation, portfolio and ledger handling, and trade execution, and evaluates their robustness under controlled system-level perturbations. All evaluations are conducted in a closed-loop historical backtesting setting on real US equity market data with identical initial conditions, enabling fair and reproducible comparisons across agents and attacks. Extensive experiments show that small perturbations at a single component can propagate through the agent decision loop and induce extreme concentration, runaway exposure, and large portfolio drawdowns across both agent types, demonstrating that current autonomous trading agents can be systematically misled at the system level. Our code is available at https://github.com/Yanlewen/TradeTrap.

Paperid: 358, https://arxiv.org/pdf/2512.02227.pdf GitHub GitHub

Abstract:
The financial market is a mission-critical playground for AI agents due to its temporal dynamics and low signal-to-noise ratio. Building an effective algorithmic trading system may require a professional team to develop and test over the years. In this paper, we propose an orchestration framework for financial agents, which aims to democratize financial intelligence to the general public. We map each component of the traditional algorithmic trading system to agents, including planner, orchestrator, alpha agents, risk agents, portfolio agents, backtest agents, execution agents, audit agents, and memory agent. We present two in-house trading examples. For the stock trading task (hourly data from 04/2024 to 12/2024), our approach achieved a return of $20.42\%$, a Sharpe ratio of 2.63, and a maximum drawdown of $-3.59\%$, while the S&P 500 index yielded a return of $15.97\%$. For the BTC trading task (minute data from 27/07/2025 to 13/08/2025), our approach achieved a return of $8.39\%$, a Sharpe ratio of $0.38$, and a maximum drawdown of $-2.80\%$, whereas the BTC price increased by $3.80\%$. Our code is available on \href{https://github.com/Open-Finance-Lab/AgenticTrading}{GitHub}.

Paperid: 359, https://arxiv.org/pdf/2512.02206.pdf GitHub

Abstract:
Sperm whales communicate in short sequences of clicks known as codas. We present WhAM (Whale Acoustics Model), the first transformer-based model capable of generating synthetic sperm whale codas from any audio prompt. WhAM is built by finetuning VampNet, a masked acoustic token model pretrained on musical audio, using 10k coda recordings collected over the past two decades. Through iterative masked token prediction, WhAM generates high-fidelity synthetic codas that preserve key acoustic features of the source recordings. We evaluate WhAM's synthetic codas using Fréchet Audio Distance and through perceptual studies with expert marine biologists. On downstream classification tasks including rhythm, social unit, and vowel classification, WhAM's learned representations achieve strong performance, despite being trained for generation rather than classification. Our code is available at https://github.com/Project-CETI/wham

Paperid: 360, https://arxiv.org/pdf/2512.02180.pdf GitHub

Abstract:
The electrocardiogram (ECG) is a key diagnostic tool in cardiovascular health. Single-lead ECG recording is integrated into both clinical-grade and consumer wearables. While self-supervised pretraining of foundation models on unlabeled ECGs improves diagnostic performance, existing approaches do not incorporate domain knowledge from clinical metadata. We introduce a novel contrastive learning approach that utilizes an established clinical risk score to adaptively weight negative pairs: clinically-guided contrastive learning. It aligns the similarities of ECG embeddings with clinically meaningful differences between subjects, with an explicit mechanism to handle missing metadata. On 12-lead ECGs from 161K patients in the MIMIC-IV dataset, we pretrain single-lead ECG foundation models at three scales, collectively called CLEF, using only routinely collected metadata without requiring per-sample ECG annotations. We evaluate CLEF on 18 clinical classification and regression tasks across 7 held-out datasets, and benchmark against 5 foundation model baselines and 3 self-supervised algorithms. When pretrained on 12-lead ECG data and tested on lead-I data, CLEF outperforms self-supervised foundation model baselines: the medium-sized CLEF achieves average AUROC improvements of at least 2.6% in classification and average reductions in MAEs of at least 3.2% in regression. Comparing with existing self-supervised learning algorithms, CLEF improves the average AUROC by at least 1.8%. Moreover, when pretrained only on lead-I data for classification tasks, CLEF performs comparably to the state-of-the-art ECGFounder, which was trained in a supervised manner. Overall, CLEF enables more accurate and scalable single-lead ECG analysis, advancing remote health monitoring. Code and pretrained CLEF models are available at: github.com/Nokia-Bell-Labs/ecg-foundation-model.

Paperid: 361, https://arxiv.org/pdf/2512.02152.pdf GitHub

Abstract:
Contrastive learning has gained popularity and pushes state-of-the-art performance across numerous large-scale benchmarks. In contrastive learning, the contrastive loss function plays a pivotal role in discerning similarities between samples through techniques such as rotation or cropping. However, this learning mechanism can also introduce information distortion from the augmented samples. This is because the trained model may develop a significant overreliance on information from samples with identical labels, while concurrently neglecting positive pairs that originate from the same initial image, especially in expansive datasets. This paper proposes a context-enriched contrastive loss function that concurrently improves learning effectiveness and addresses the information distortion by encompassing two convergence targets. The first component, which is notably sensitive to label contrast, differentiates between features of identical and distinct classes which boosts the contrastive training efficiency. Meanwhile, the second component draws closer the augmented samples from the same source image and distances all other samples. We evaluate the proposed approach on image classification tasks, which are among the most widely accepted 8 recognition large-scale benchmark datasets: CIFAR10, CIFAR100, Caltech-101, Caltech-256, ImageNet, BiasedMNIST, UTKFace, and CelebA datasets. The experimental results demonstrate that the proposed method achieves improvements over 16 state-of-the-art contrastive learning methods in terms of both generalization performance and learning convergence speed. Interestingly, our technique stands out in addressing systematic distortion tasks. It demonstrates a 22.9% improvement compared to original contrastive loss functions in the downstream BiasedMNIST dataset, highlighting its promise for more efficient and equitable downstream training.

Paperid: 362, https://arxiv.org/pdf/2512.02070.pdf GitHub

Abstract:
Long-term time series forecasting (LTSF) is a critical task in computational intelligence. While Transformer-based models effectively capture long-range dependencies, they often suffer from quadratic complexity and overfitting due to data sparsity. Conversely, efficient linear models struggle to depict complex non-linear local dynamics. Furthermore, existing multi-scale frameworks typically rely on average pooling, which acts as a non-ideal low-pass filter, leading to spectral aliasing and the irreversible loss of high-frequency transients. In response, this paper proposes DPWMixer, a computationally efficient Dual-Path architecture. The framework is built upon a Lossless Haar Wavelet Pyramid that replaces traditional pooling, utilizing orthogonal decomposition to explicitly disentangle trends and local fluctuations without information loss. To process these components, we design a Dual-Path Trend Mixer that integrates a global linear mapping for macro-trend anchoring and a flexible patch-based MLP-Mixer for micro-dynamic evolution. Finally, An adaptive multi-scale fusion module then integrates predictions from diverse scales, weighted by channel stationarity to optimize synthesis. Extensive experiments on eight public benchmarks demonstrate that our method achieves a consistent improvement over state-of-the-art baselines. The code is available at https://github.com/hit636/DPWMixer.

Paperid: 363, https://arxiv.org/pdf/2512.02062.pdf GitHub

Abstract:
Deep learning models are used in safety-critical tasks such as automated driving and face recognition. However, small perturbations in the model input can significantly change the predictions. Adversarial attacks are used to identify small perturbations that can lead to misclassifications. More powerful black-box adversarial attacks are required to develop more effective defenses. A promising approach to black-box adversarial attacks is to repeat the process of extracting a specific image area and changing the perturbations added to it. Existing attacks adopt simple rectangles as the areas where perturbations are changed in a single iteration. We propose applying superpixels instead, which achieve a good balance between color variance and compactness. We also propose a new search method, versatile search, and a novel attack method, Superpixel Attack, which applies superpixels and performs versatile search. Superpixel Attack improves attack success rates by an average of 2.10% compared with existing attacks. Most models used in this study are robust against adversarial attacks, and this improvement is significant for black-box adversarial attacks. The code is avilable at https://github.com/oe1307/SuperpixelAttack.git.

Paperid: 364, https://arxiv.org/pdf/2512.02021.pdf GitHub

Abstract:
Conventional database architectures often secure local consistency by discarding information, entangling correctness with loss. We introduce the Functorial-Categorical Database (FCDb), which models data operations as morphisms in a layered functor category and establishes a Complete Preserving Family (CPF) of projections spanning content invariance (CAS), capability, and ownership, with optional observational projections for local order (B+Tree), temporal history (append-only/LSM), and adjacency (Graph). We identify a minimal kernel (F_core = Own o Cap o CAS) that preserves information and collapses non-commutativity to the ethical grant/revoke boundary. Under adjoint lifts and a fibred structure, operational pairs commute in the categorical limit while ownership integrity and capability constraints are maintained. The framework connects to information geometry via projection interpretations and supports empirical validation without discarding semantic, temporal, or relational entropy.

Paperid: 365, https://arxiv.org/pdf/2512.02017.pdf GitHub

Abstract:
Today, people can easily record memorable moments, ranging from concerts, sports events, lectures, family gatherings, and birthday parties with multiple consumer cameras. However, synchronizing these cross-camera streams remains challenging. Existing methods assume controlled settings, specific targets, manual correction, or costly hardware. We present VisualSync, an optimization framework based on multi-view dynamics that aligns unposed, unsynchronized videos at millisecond accuracy. Our key insight is that any moving 3D point, when co-visible in two cameras, obeys epipolar constraints once properly synchronized. To exploit this, VisualSync leverages off-the-shelf 3D reconstruction, feature matching, and dense tracking to extract tracklets, relative poses, and cross-view correspondences. It then jointly minimizes the epipolar error to estimate each camera's time offset. Experiments on four diverse, challenging datasets show that VisualSync outperforms baseline methods, achieving an median synchronization error below 50 ms.

Paperid: 366, https://arxiv.org/pdf/2512.02010.pdf GitHub

Abstract:
As large language models have grown larger, low-precision numerical formats such as NVFP4 have become increasingly popular due to the speed and memory benefits they provide. However, to accelerate computation with NVFP4, all matrix multiplication operands--weights and activations in the forward pass, and weights, activations, and gradients in the backward pass--must be quantized to NVFP4, often leading to divergence during training and performance degradation during inference. To address this issue, in this work we introduce Four Over Six (4/6), a modification to the NVFP4 quantization algorithm that evaluates two potential scale factors for each block of values. Unlike integer formats, floating-point formats such as FP4 have the most quantization error on near-maximal values in each block, which we find to be primarily responsible for downstream performance degradation. We find that for some blocks, scaling to smaller FP4 values makes the distribution of representable values more uniform, improving representation of near-maximal values. Importantly, 4/6 can be implemented efficiently on NVIDIA Blackwell GPUs, making it viable to use while training LLMs with NVFP4. In pre-training experiments with transformer and hybrid model architectures, we find that 4/6 prevents divergence in several cases, bringing training loss significantly closer to BF16 compared to models trained with current state-of-the-art NVFP4 training recipes. We also find that 4/6 can be easily incorporated into many different post-training quantization methods and generally improves downstream accuracy. We hope this inspires future work in training and deploying models with NVFP4. Our code is available at http://github.com/mit-han-lab/fouroversix.

Paperid: 367, https://arxiv.org/pdf/2512.02009.pdf GitHub GitHub

Abstract:
The field of 360-degree omnidirectional understanding has been receiving increasing attention for advancing spatial intelligence. However, the lack of large-scale and diverse data remains a major limitation. In this work, we propose AirSim360, a simulation platform for omnidirectional data from aerial viewpoints, enabling wide-ranging scene sampling with drones. Specifically, AirSim360 focuses on three key aspects: a render-aligned data and labeling paradigm for pixel-level geometric, semantic, and entity-level understanding; an interactive pedestrian-aware system for modeling human behavior; and an automated trajectory generation paradigm to support navigation tasks. Furthermore, we collect more than 60K panoramic samples and conduct extensive experiments across various tasks to demonstrate the effectiveness of our simulator. Unlike existing simulators, our work is the first to systematically model the 4D real world under an omnidirectional setting. The entire platform, including the toolkit, plugins, and collected datasets, will be made publicly available at https://insta360-research-team.github.io/AirSim360-website.

Paperid: 368, https://arxiv.org/pdf/2512.02006.pdf GitHub

Abstract:
Multi-view camera systems enable rich observations of complex real-world scenes, and understanding dynamic objects in multi-view settings has become central to various applications. In this work, we present MV-TAP, a novel point tracker that tracks points across multi-view videos of dynamic scenes by leveraging cross-view information. MV-TAP utilizes camera geometry and a cross-view attention mechanism to aggregate spatio-temporal information across views, enabling more complete and reliable trajectory estimation in multi-view videos. To support this task, we construct a large-scale synthetic training dataset and real-world evaluation sets tailored for multi-view tracking. Extensive experiments demonstrate that MV-TAP outperforms existing point-tracking methods on challenging benchmarks, establishing an effective baseline for advancing research in multi-view point tracking.

Paperid: 369, https://arxiv.org/pdf/2512.02005.pdf GitHub

Abstract:
We introduce Audio-Visual Affordance Grounding (AV-AG), a new task that segments object interaction regions from action sounds. Unlike existing approaches that rely on textual instructions or demonstration videos, which often limited by ambiguity or occlusion, audio provides real-time, semantically rich, and visually independent cues for affordance grounding, enabling more intuitive understanding of interaction regions. To support this task, we construct the first AV-AG dataset, comprising a large collection of action sounds, object images, and pixel-level affordance annotations. The dataset also includes an unseen subset to evaluate zero-shot generalization. Furthermore, we propose AVAGFormer, a model equipped with a semantic-conditioned cross-modal mixer and a dual-head decoder that effectively fuses audio and visual signals for mask prediction. Experiments show that AVAGFormer achieves state-of-the-art performance on AV-AG, surpassing baselines from related tasks. Comprehensive analyses highlight the distinctions between AV-AG and AVS, the benefits of end-to-end modeling, and the contribution of each component. Code and dataset have been released on https://jscslld.github.io/AVAGFormer/.

Paperid: 370, https://arxiv.org/pdf/2512.01980.pdf GitHub

Abstract:
Low-rank approximation methods such as singular value decomposition (SVD) and its variants (e.g., Fisher-weighted SVD, Activation SVD) have recently emerged as effective tools for neural network compression. In this setting, decomposition acts as a "surgical" intervention, followed by fine-tuning that serves as "rehab" to recover accuracy. Inspired by prehabilitation in surgery, we introduce a pre-compression fine-tuning stage, Low-Rank Prehab, that explicitly encourages low-rank structure in weight matrices while preserving task performance. By conditioning the model before SVD, Prehab steers weights toward spectrally compact regions of the parameter space, enabling smoother low-rank approximation and improved recovery. Experiments on large language models (LLMs) and other Transformer-based architectures, including Vision Transformers (ViTs), show that Prehab substantially reduces the immediate accuracy drop after compression and consistently improves post-finetuning performance. Across a wide range of compression ratios, our method outperforms state-of-the-art SVD-based techniques such as SVD-LLM, highlighting the importance of preparing models for compression rather than only improving the compression and recovery stages. Source code is available at https://github.com/niqretnuh/PREHAB-SVD

Paperid: 371, https://arxiv.org/pdf/2512.01970.pdf GitHub

Abstract:
The mechanism by which RL contributes to reasoning capabilities-whether it incentivizes the synthesis of new skills or merely amplifies existing behaviors-remains a subject of intense debate. In this work, we investigate this question through the lens of Complementary Reasoning, a complex task that requires integrating internal parametric knowledge with external contextual information. Using a controlled synthetic dataset of human biographies, we strictly decouple this ability into two atomic skills: Parametric Reasoning (relying on internal knowledge) and Contextual Reasoning (depending on external information). To rigorously assess capability boundaries, we evaluate generalization across three distinct levels of difficulty: I.I.D., Composition, and Zero-shot settings. We find that while SFT is sufficient for in-distribution performance, it struggles with O.O.D. generalization, particularly in Zero-shot settings where relational combinations are novel. Crucially, we identify the SFT Generalization Paradox: Models supervised solely on the composite task achieve near-perfect in-distribution accuracy but collapse on out-of-distribution generalization, indicating their reliance on rote memorization of path shortcuts. In contrast, we find that RL acts as a reasoning synthesizer rather than a probability amplifier. However, we uncover a strict atomic prerequisite: RL can only synthesize these complex strategies if the base model has first mastered the independent atomic skills (Parametric and Contextual) via SFT. These findings challenge the view of RL as a mere amplifier, suggesting that given sufficient atomic foundations, RL can actively synthesize complex reasoning strategies from learned primitives without explicit supervision on such complex strategies. This indicates that decoupled atomic training followed by RL offers a scalable path to generalization for complex reasoning tasks.

Paperid: 372, https://arxiv.org/pdf/2512.01949.pdf GitHub

Abstract:
The rapid growth of visual tokens in multimodal large language models (MLLMs) leads to excessive memory consumption and inference latency, especially when handling high-resolution images and videos. Token pruning is a technique used to mitigate this issue by removing redundancy, but existing methods often ignore relevance to the user query or suffer from the limitations of attention mechanisms, reducing their adaptability and effectiveness. To address these challenges, we propose Script, a plug-and-play pruning method that requires no retraining and generalizes across diverse MLLMs. Script comprises two modules: a graph-structured pruning module that removes visually redundant tokens, and a query-conditioned semantic pruning module that preserves query-relevant visual information. Together, they enhance performance on multimodal tasks. Experiments on fourteen benchmarks across image and video understanding tasks show that Script consistently achieves higher model efficiency and predictive accuracy compared to existing pruning methods. On LLaVA-NeXT-7B, it achieves up to 6.8x prefill speedup and 10x FLOP reduction, while retaining 96.88% of the original performance.

Paperid: 373, https://arxiv.org/pdf/2512.01913.pdf GitHub

Abstract:
Medical image registration drives quantitative analysis across organs, modalities, and patient populations. Recent deep learning methods often combine low-level "trend-driven" computational blocks from computer vision, such as large-kernel CNNs, Transformers, and state-space models, with high-level registration-specific designs like motion pyramids, correlation layers, and iterative refinement. Yet, their relative contributions remain unclear and entangled. This raises a central question: should future advances in registration focus on importing generic architectural trends or on refining domain-specific design principles? Through a modular framework spanning brain, lung, cardiac, and abdominal registration, we systematically disentangle the influence of these two paradigms. Our evaluation reveals that low-level "trend-driven" computational blocks offer only marginal or inconsistent gains, while high-level registration-specific designs consistently deliver more accurate, smoother, and more robust deformations. These domain priors significantly elevate the performance of a standard U-Net baseline, far more than variants incorporating "trend-driven" blocks, achieving an average relative improvement of $\sim3\%$. All models and experiments are released within a transparent, modular benchmark that enables plug-and-play comparison for new architectures and registration tasks (https://github.com/BailiangJ/rethink-reg). This dynamic and extensible platform establishes a common ground for reproducible and fair evaluation, inviting the community to isolate genuine methodological contributions from domain priors. Our findings advocate a shift in research emphasis: from following architectural trends to embracing domain-specific design principles as the true drivers of progress in learning-based medical image registration.

Paperid: 374, https://arxiv.org/pdf/2512.01885.pdf GitHub

Abstract:
Tracking cells in time-lapse videos is an essential technique for monitoring cell population dynamics at a single-cell level. Current methods for cell tracking are developed on videos with mostly single, constant signals and do not detect pivotal events such as cell death. Here, we present TransientTrack, a deep learning-based framework for cell tracking in multi-channel microscopy video data with transient fluorescent signals that fluctuate over time following processes such as the circadian rhythm of cells. By identifying key cellular events - mitosis (cell division) and apoptosis (cell death) our method allows us to build complete trajectories, including cell lineage information. TransientTrack is lightweight and performs matching on cell detection embeddings directly, without the need for quantification of tracking-specific cell features. Furthermore, our approach integrates Transformer Networks, multi-stage matching using all detection boxes, and the interpolation of missing tracklets with the Kalman Filter. This unified framework achieves strong performance across diverse conditions, effectively tracking cells and capturing cell division and death. We demonstrate the use of TransientTrack in an analysis of the efficacy of a chemotherapeutic drug at a single-cell level. The proposed framework could further advance quantitative studies of cancer cell dynamics, enabling detailed characterization of treatment response and resistance mechanisms. The code is available at https://github.com/bozeklab/TransientTrack.

Paperid: 375, https://arxiv.org/pdf/2512.01853.pdf GitHub

Abstract:
Intelligent sports video analysis demands a comprehensive understanding of temporal context, from micro-level actions to macro-level game strategies. Existing end-to-end models often struggle with this temporal hierarchy, offering solutions that lack generalization, incur high development costs for new tasks, and suffer from poor interpretability. To overcome these limitations, we propose a reconfigurable Multi-Agent System (MAS) as a foundational framework for sports video understanding. In our system, each agent functions as a distinct "cognitive tool" specializing in a specific aspect of analysis. The system's architecture is not confined to a single temporal dimension or task. By leveraging iterative invocation and flexible composition of these agents, our framework can construct adaptive pipelines for both short-term analytic reasoning (e.g., Rally QA) and long-term generative summarization (e.g., match summaries). We demonstrate the adaptability of this framework using two representative tasks in badminton analysis, showcasing its ability to bridge fine-grained event detection and global semantic organization. This work presents a paradigm shift towards a flexible, scalable, and interpretable system for robust, cross-task sports video intelligence. The project homepage is available at https://aiden1020.github.io/COACH-project-page

Paperid: 376, https://arxiv.org/pdf/2512.01852.pdf GitHub

Abstract:
Large language models (LLMs) are increasingly deployed in multilingual applications but often generate plausible yet incorrect or misleading outputs, known as hallucinations. While hallucination detection has been studied extensively in English, under-resourced Indian languages remain largely unexplored. We present BHRAM-IL, a benchmark for hallucination recognition and assessment in multiple Indian languages, covering Hindi, Gujarati, Marathi, Odia, along with English. The benchmark comprises 36,047 curated questions across nine categories spanning factual, numerical, reasoning, and linguistic tasks. We evaluate 14 state-of-the-art multilingual LLMs on a benchmark subset of 10,265 questions, analyzing cross-lingual and factual hallucinations across languages, models, scales, categories, and domains using category-specific metrics normalized to (0,1) range. Aggregation over all categories and models yields a primary score of 0.23 and a language-corrected fuzzy score of 0.385, demonstrating the usefulness of BHRAM-IL for hallucination-focused evaluation. The dataset, and the code for generation and evaluation are available on GitHub (https://github.com/sambhashana/BHRAM-IL/) and HuggingFace (https://huggingface.co/datasets/sambhashana/BHRAM-IL/) to support future research in multilingual hallucination detection and mitigation.

Paperid: 377, https://arxiv.org/pdf/2512.01850.pdf GitHub

Abstract:
Point cloud registration aligns multiple unposed point clouds into a common frame, and is a core step for 3D reconstruction and robot localization. In this work, we cast registration as conditional generation: a learned continuous, point-wise velocity field transports noisy points to a registered scene, from which the pose of each view is recovered. Unlike previous methods that conduct correspondence matching to estimate the transformation between a pair of point clouds and then optimize the pairwise transformations to realize multi-view registration, our model directly generates the registered point cloud. With a lightweight local feature extractor and test-time rigidity enforcement, our approach achieves state-of-the-art results on pairwise and multi-view registration benchmarks, particularly with low overlap, and generalizes across scales and sensor modalities. It further supports downstream tasks including relocalization, multi-robot SLAM, and multi-session map merging. Source code available at: https://github.com/PRBonn/RAP.

Paperid: 378, https://arxiv.org/pdf/2512.01843.pdf GitHub GitHub

Abstract:
Driven by the growing capacity and training scale, Text-to-Video (T2V) generation models have recently achieved substantial progress in video quality, length, and instruction-following capability. However, whether these models can understand physics and generate physically plausible videos remains a question. While Vision-Language Models (VLMs) have been widely used as general-purpose evaluators in various applications, they struggle to identify the physically impossible content from generated videos. To investigate this issue, we construct a \textbf{PID} (\textbf{P}hysical \textbf{I}mplausibility \textbf{D}etection) dataset, which consists of a \textit{test split} of 500 manually annotated videos and a \textit{train split} of 2,588 paired videos, where each implausible video is generated by carefully rewriting the caption of its corresponding real-world video to induce T2V models producing physically implausible content. With the constructed dataset, we introduce a lightweight fine-tuning approach, enabling VLMs to not only detect physically implausible events but also generate textual explanations on the violated physical principles. Taking the fine-tuned VLM as a physical plausibility detector and explainer, namely \textbf{PhyDetEx}, we benchmark a series of state-of-the-art T2V models to assess their adherence to physical laws. Our findings show that although recent T2V models have made notable progress toward generating physically plausible content, understanding and adhering to physical laws remains a challenging issue, especially for open-source models. Our dataset, training code, and checkpoints are available at \href{https://github.com/Zeqing-Wang/PhyDetEx}{https://github.com/Zeqing-Wang/PhyDetEx}.

Paperid: 379, https://arxiv.org/pdf/2512.01827.pdf GitHub GitHub

Abstract:
Causal thinking enables humans to understand not just what is seen, but why it happens. To replicate this capability in modern AI systems, we introduce the task of visual causal discovery. It requires models to infer cause-and-effect relations among visual entities across diverse scenarios instead of merely perceiving their presence. To this end, we first construct the Visual Causal Graph dataset (VCG-32K), a large-scale collection of over 32,000 images annotated with entity-level causal graphs, and further develop CauSight, a novel vision-language model to perform visual causal discovery through causally aware reasoning. Our training recipe integrates three components: (1) training data curation from VCG-32K, (2) Tree-of-Causal-Thought (ToCT) for synthesizing reasoning trajectories, and (3) reinforcement learning with a designed causal reward to refine the reasoning policy. Experiments show that CauSight outperforms GPT-4.1 on visual causal discovery, achieving over a threefold performance boost (21% absolute gain). Our code, model, and dataset are fully open-sourced at project page: https://github.com/OpenCausaLab/CauSight.

Paperid: 380, https://arxiv.org/pdf/2512.01809.pdf GitHub

Abstract:
Generative models, like flows and diffusions, have recently emerged as popular and efficacious policy parameterizations in robotics. There has been much speculation as to the factors underlying their successes, ranging from capturing multi-modal action distribution to expressing more complex behaviors. In this work, we perform a comprehensive evaluation of popular generative control policies (GCPs) on common behavior cloning (BC) benchmarks. We find that GCPs do not owe their success to their ability to capture multi-modality or to express more complex observation-to-action mappings. Instead, we find that their advantage stems from iterative computation, as long as intermediate steps are supervised during training and this supervision is paired with a suitable level of stochasticity. As a validation of our findings, we show that a minimum iterative policy (MIP), a lightweight two-step regression-based policy, essentially matches the performance of flow GCPs, and often outperforms distilled shortcut models. Our results suggest that the distribution-fitting component of GCPs is less salient than commonly believed, and point toward new design spaces focusing solely on control performance. Project page: https://simchowitzlabpublic.github.io/much-ado-about-noising-project/

Paperid: 381, https://arxiv.org/pdf/2512.01789.pdf GitHub

Abstract:
In this paper, we introduce SAM3-UNet, a simplified variant of Segment Anything Model 3 (SAM3), designed to adapt SAM3 for downstream tasks at a low cost. Our SAM3-UNet consists of three components: a SAM3 image encoder, a simple adapter for parameter-efficient fine-tuning, and a lightweight U-Net-style decoder. Preliminary experiments on multiple tasks, such as mirror detection and salient object detection, demonstrate that the proposed SAM3-UNet outperforms the prior SAM2-UNet and other state-of-the-art methods, while requiring less than 6 GB of GPU memory during training with a batch size of 12. The code is publicly available at https://github.com/WZH0120/SAM3-UNet.

Paperid: 382, https://arxiv.org/pdf/2512.01753.pdf GitHub

Abstract:
Multi-sensor Simultaneous Localization and Mapping (SLAM) is essential for Unmanned Aerial Vehicles (UAVs) performing agricultural tasks such as spraying, surveying, and inspection. However, real-world, multi-modal agricultural UAV datasets that enable research on robust operation remain scarce. To address this gap, we present AgriLiRa4D, a multi-modal UAV dataset designed for challenging outdoor agricultural environments. AgriLiRa4D spans three representative farmland types-flat, hilly, and terraced-and includes both boundary and coverage operation modes, resulting in six flight sequence groups. The dataset provides high-accuracy ground-truth trajectories from a Fiber Optic Inertial Navigation System with Real-Time Kinematic capability (FINS_RTK), along with synchronized measurements from a 3D LiDAR, a 4D Radar, and an Inertial Measurement Unit (IMU), accompanied by complete intrinsic and extrinsic calibrations. Leveraging its comprehensive sensor suite and diverse real-world scenarios, AgriLiRa4D supports diverse SLAM and localization studies and enables rigorous robustness evaluation against low-texture crops, repetitive patterns, dynamic vegetation, and other challenges of real agricultural environments. To further demonstrate its utility, we benchmark four state-of-the-art multi-sensor SLAM algorithms across different sensor combinations, highlighting the difficulty of the proposed sequences and the necessity of multi-modal approaches for reliable UAV localization. By filling a critical gap in agricultural SLAM datasets, AgriLiRa4D provides a valuable benchmark for the research community and contributes to advancing autonomous navigation technologies for agricultural UAVs. The dataset can be downloaded from: https://zhan994.github.io/AgriLiRa4D.

Paperid: 383, https://arxiv.org/pdf/2512.01723.pdf GitHub

Abstract:
Modeling historical events poses fundamental challenges for machine learning: extreme data scarcity (N << 100), heterogeneous and noisy measurements, missing counterfactuals, and the requirement for human interpretable explanations. We present HistoricalML, a probabilistic neuro-symbolic framework that addresses these challenges through principled integration of (1) Bayesian uncertainty quantification to separate epistemic from aleatoric uncertainty, (2) structural causal models for counterfactual reasoning under confounding, (3) cooperative game theory (Shapley values) for fair allocation modeling, and (4) attention based neural architectures for context dependent factor weighting. We provide theoretical analysis showing that our approach achieves consistent estimation in the sparse data regime when strong priors from domain knowledge are available, and that Shapley based allocation satisfies axiomatic fairness guarantees that pure regression approaches cannot provide. We instantiate the framework on two historical case studies: the 19th century partition of Africa (N = 7 colonial powers) and the Second Punic War (N = 2 factions). Our model identifies Germany's +107.9 percent discrepancy as a quantifiable structural tension preceding World War I, with tension factor 36.43 and 0.79 naval arms race correlation. For the Punic Wars, Monte Carlo battle simulations achieve a 57.3 percent win probability for Carthage at Cannae and 57.8 percent for Rome at Zama, aligning with historical outcomes. Counterfactual analysis reveals that Carthaginian political support (support score 6.4 vs Napoleon's 7.1), rather than military capability, was the decisive factor.

Paperid: 384, https://arxiv.org/pdf/2512.01686.pdf GitHub

Abstract:
Current story visualization methods tend to position subjects solely by text and face challenges in maintaining artistic consistency. To address these limitations, we introduce DreamingComics, a layout-aware story visualization framework. We build upon a pretrained video diffusion-transformer (DiT) model, leveraging its spatiotemporal priors to enhance identity and style consistency. For layout-based position control, we propose RegionalRoPE, a region-aware positional encoding scheme that re-indexes embeddings based on the target layout. Additionally, we introduce a masked condition loss to further constrain each subject's visual features to their designated region. To infer layouts from natural language scripts, we integrate an LLM-based layout generator trained to produce comic-style layouts, enabling flexible and controllable layout conditioning. We present a comprehensive evaluation of our approach, showing a 29.2% increase in character consistency and a 36.2% increase in style similarity compared to previous methods, while displaying high spatial accuracy. Our project page is available at https://yj7082126.github.io/dreamingcomics/

Paperid: 385, https://arxiv.org/pdf/2512.01677.pdf GitHub

Abstract:
Generating realistic hand-object interactions (HOI) videos is a significant challenge due to the difficulty of modeling physical constraints (e.g., contact and occlusion between hands and manipulated objects). Current methods utilize HOI representation as an auxiliary generative objective to guide video synthesis. However, there is a dilemma between 2D and 3D representations that cannot simultaneously guarantee scalability and interaction fidelity. To address this limitation, we propose a structure and contact-aware representation that captures hand-object contact, hand-object occlusion, and holistic structure context without 3D annotations. This interaction-oriented and scalable supervision signal enables the model to learn fine-grained interaction physics and generalize to open-world scenarios. To fully exploit the proposed representation, we introduce a joint-generation paradigm with a share-and-specialization strategy that generates interaction-oriented representations and videos. Extensive experiments demonstrate that our method outperforms state-of-the-art methods on two real-world datasets in generating physics-realistic and temporally coherent HOI videos. Furthermore, our approach exhibits strong generalization to challenging open-world scenarios, highlighting the benefit of our scalable design. Our project page is https://hgzn258.github.io/SCAR/.

Paperid: 386, https://arxiv.org/pdf/2512.01661.pdf GitHub

Abstract:
Ensuring LLM reliability requires not only solving complex problems but also recognizing when a problem is unsolvable. Current models often struggle to distinguish objective unsolvability (inherent contradictions in the problem) from subjective capability limitations (problems beyond the model's competence), which leads to hallucinations and overconfidence. To address this, we propose UnsolvableQA and UnsolvableRL to solve feasible problems, detect inherent contradictions, and prudently refuse tasks beyond capability. Specifically, we construct UnsolvableQA, a dataset of paired solvable and unsolvable instances derived via a dual-track methodology: programmatic generation for logic puzzles and a novel "Reverse Construction" method that injects contradictions into valid reasoning chains for mathematics. Building on this dataset, we introduce UnsolvableRL, a reinforcement learning framework with three reward components jointly accounting for accuracy, unsolvability, and difficulty. Empirical results show that our approach achieves near-perfect unsolvability detection while also improving accuracy on solvable tasks. Crucially, we identify Capability Collapse, demonstrating that explicit exposure to unsolvable data is indispensable for preventing models from becoming systematically overconfident. Our code and data are available at https://github.com/sfasfaffa/unsolvableQA.

Paperid: 387, https://arxiv.org/pdf/2512.01643.pdf GitHub

Abstract:
Test-Time Training (TTT) has recently emerged as a promising direction for efficient sequence modeling. TTT reformulates attention operation as an online learning problem, constructing a compact inner model from key-value pairs at test time. This reformulation opens a rich and flexible design space while achieving linear computational complexity. However, crafting a powerful visual TTT design remains challenging: fundamental choices for the inner module and inner training lack comprehensive understanding and practical guidelines. To bridge this critical gap, in this paper, we present a systematic empirical study of TTT designs for visual sequence modeling. From a series of experiments and analyses, we distill six practical insights that establish design principles for effective visual TTT and illuminate paths for future improvement. These findings culminate in the Vision Test-Time Training (ViT$^3$) model, a pure TTT architecture that achieves linear complexity and parallelizable computation. We evaluate ViT$^3$ across diverse visual tasks, including image classification, image generation, object detection, and semantic segmentation. Results show that ViT$^3$ consistently matches or outperforms advanced linear-complexity models (e.g., Mamba and linear attention variants) and effectively narrows the gap to highly optimized vision Transformers. We hope this study and the ViT$^3$ baseline can facilitate future work on visual TTT models. Code is available at https://github.com/LeapLabTHU/ViTTT.

Paperid: 388, https://arxiv.org/pdf/2512.01630.pdf GitHub

Abstract:
Software supply chain attacks have revealed blind spots in existing SCA tools, which are often limited to a single ecosystem and assess either software artifacts or community activity in isolation. This fragmentation across tools and ecosystems forces developers to manually reconcile scattered data, undermining risk assessments. We present Package Dashboard, a cross-ecosystem framework that provides a unified platform for supply chain analysis, enabling a holistic, dual-perspective risk assessment by integrating package metadata, vulnerability information, and upstream community health metrics. By combining dependency resolution with repository analysis, it reduces cognitive load and improves traceability. Demonstrating the framework's versatility, a large-scale study of 374,000 packages across five Linux distributions shows its ability to uncover not only conventional vulnerabilities and license conflicts but also overlooked risks such as archived or inaccessible repositories. Ultimately, Package Dashboard provides a unified view of risk, equipping developers and DevSecOps engineers with actionable insights to strengthen the transparency, trustworthiness, and traceability of open-source ecosystems. Package Dashboard is publicly available at https://github.com/n19htfall/PackageDashboard, and a demonstration video can be found at https://youtu.be/y9ncftP8KPQ. Besides, the online version is available at https://pkgdash.osslab-pku.org.

Paperid: 389, https://arxiv.org/pdf/2512.01603.pdf GitHub

Abstract:
Spoken Language Understanding (SLU), which aims to extract user semantics to execute downstream tasks, is a crucial component of task-oriented dialog systems. Existing SLU datasets generally lack sufficient diversity and complexity, and there is an absence of a unified benchmark for the latest Large Language Models (LLMs) and Large Audio Language Models (LALMs). This work introduces MAC-SLU, a novel Multi-Intent Automotive Cabin Spoken Language Understanding Dataset, which increases the difficulty of the SLU task by incorporating authentic and complex multi-intent data. Based on MAC-SLU, we conducted a comprehensive benchmark of leading open-source LLMs and LALMs, covering methods like in-context learning, supervised fine-tuning (SFT), and end-to-end (E2E) and pipeline paradigms. Our experiments show that while LLMs and LALMs have the potential to complete SLU tasks through in-context learning, their performance still lags significantly behind SFT. Meanwhile, E2E LALMs demonstrate performance comparable to pipeline approaches and effectively avoid error propagation from speech recognition. Code\footnote{https://github.com/Gatsby-web/MAC\_SLU} and datasets\footnote{huggingface.co/datasets/Gatsby1984/MAC\_SLU} are released publicly.

Paperid: 390, https://arxiv.org/pdf/2512.01589.pdf GitHub

Abstract:
Abscesses in the head and neck represent an acute infectious process that can potentially lead to sepsis or mortality if not diagnosed and managed promptly. Accurate detection and delineation of these lesions on imaging are essential for diagnosis, treatment planning, and surgical intervention. In this study, we introduce AbscessHeNe, a curated and comprehensively annotated dataset comprising 4,926 contrast-enhanced CT slices with clinically confirmed head and neck abscesses. The dataset is designed to facilitate the development of robust semantic segmentation models that can accurately delineate abscess boundaries and evaluate deep neck space involvement, thereby supporting informed clinical decision-making. To establish performance baselines, we evaluate several state-of-the-art segmentation architectures, including CNN, Transformer, and Mamba-based models. The highest-performing model achieved a Dice Similarity Coefficient of 0.39, Intersection-over-Union of 0.27, and Normalized Surface Distance of 0.67, indicating the challenges of this task and the need for further research. Beyond segmentation, AbscessHeNe is structured for future applications in content-based multimedia indexing and case-based retrieval. Each CT scan is linked with pixel-level annotations and clinical metadata, providing a foundation for building intelligent retrieval systems and supporting knowledge-driven clinical workflows. The dataset will be made publicly available at https://github.com/drthaodao3101/AbscessHeNe.git.

Paperid: 391, https://arxiv.org/pdf/2512.01568.pdf GitHub

Abstract:
We investigate whether Large Language Models (LLMs) exhibit altruistic tendencies, and critically, whether their implicit associations and self-reports predict actual altruistic behavior. Using a multi-method approach inspired by human social psychology, we tested 24 frontier LLMs across three paradigms: (1) an Implicit Association Test (IAT) measuring implicit altruism bias, (2) a forced binary choice task measuring behavioral altruism, and (3) a self-assessment scale measuring explicit altruism beliefs. Our key findings are: (1) All models show strong implicit pro-altruism bias (mean IAT = 0.87, p < .0001), confirming models "know" altruism is good. (2) Models behave more altruistically than chance (65.6% vs. 50%, p < .0001), but with substantial variation (48-85%). (3) Implicit associations do not predict behavior (r = .22, p = .29). (4) Most critically, models systematically overestimate their own altruism, claiming 77.5% altruism while acting at 65.6% (p < .0001, Cohen's d = 1.08). This "virtue signaling gap" affects 75% of models tested. Based on these findings, we recommend the Calibration Gap (the discrepancy between self-reported and behavioral values) as a standardized alignment metric. Well-calibrated models are more predictable and behaviorally consistent; only 12.5% of models achieve the ideal combination of high prosocial behavior and accurate self-knowledge.

Paperid: 392, https://arxiv.org/pdf/2512.01563.pdf GitHub

Abstract:
Head and neck masses are space-occupying lesions that can compress the airway and esophagus and may affect nerves and blood vessels. Available public datasets primarily focus on malignant lesions and often overlook other space-occupying conditions in this region. To address this gap, we introduce MasHeNe, an initial dataset of 3,779 contrast-enhanced CT slices that includes both tumors and cysts with pixel-level annotations. We also establish a benchmark using standard segmentation baselines and report common metrics to enable fair comparison. In addition, we propose the Windowing-Enhanced Mamba with Frequency integration (WEMF) model. WEMF applies tri-window enhancement to enrich the input appearance before feature extraction. It further uses multi-frequency attention to fuse information across skip connections within a U-shaped Mamba backbone. On MasHeNe, WEMF attains the best performance among evaluated methods, with a Dice of 70.45%, IoU of 66.89%, NSD of 72.33%, and HD95 of 5.12 mm. This model indicates stable and strong results on this challenging task. MasHeNe provides a benchmark for head-and-neck mass segmentation beyond malignancy-only datasets. The observed error patterns also suggest that this task remains challenging and requires further research. Our dataset and code are available at https://github.com/drthaodao3101/MasHeNe.git.

Paperid: 393, https://arxiv.org/pdf/2512.01540.pdf GitHub

Abstract:
3D reconstruction from multi-view images is a core challenge in computer vision. Recently, feed-forward methods have emerged as efficient and robust alternatives to traditional per-scene optimization techniques. Among them, state-of-the-art models like the Visual Geometry Grounding Transformer (VGGT) leverage full self-attention over all image tokens to capture global relationships. However, this approach suffers from poor scalability due to the quadratic complexity of self-attention and the large number of tokens generated in long image sequences. In this work, we introduce FlashVGGT, an efficient alternative that addresses this bottleneck through a descriptor-based attention mechanism. Instead of applying dense global attention across all tokens, FlashVGGT compresses spatial information from each frame into a compact set of descriptor tokens. Global attention is then computed as cross-attention between the full set of image tokens and this smaller descriptor set, significantly reducing computational overhead. Moreover, the compactness of the descriptors enables online inference over long sequences via a chunk-recursive mechanism that reuses cached descriptors from previous chunks. Experimental results show that FlashVGGT achieves reconstruction accuracy competitive with VGGT while reducing inference time to just 9.3% of VGGT for 1,000 images, and scaling efficiently to sequences exceeding 3,000 images. Our project page is available at https://wzpscott.github.io/flashvggt_page/.

Paperid: 394, https://arxiv.org/pdf/2512.01528.pdf GitHub

Abstract:
We revisit the notion of Feedback Integrators introduced by D. E. Chang in 2016. Feedback integrators allow for numerically integrating dynamical systems on manifold while preserving the first integrals of the system. However, its performance was stated and proved in an asymptotic manner, which left a gap between its empirical success and theoretical understandings. In response, we prove preservation of first integrals over entire integration region up to arbitrarily small deviation under Feedback Integrator framework. Furthermore, we propose an adaptive gain selection scheme that significantly improves the performance. Numerical demonstrations are conducted on free rigid body motion in SO(3), the Kepler problem, and a perturbed Kepler problem with rotational symmetry. All demonstration codes are available at: https://github.com/johnbae1901/Feedback-Integrator.

Paperid: 395, https://arxiv.org/pdf/2512.01519.pdf GitHub

Abstract:
Despite rapid advances in molecular and materials machine learning, most models still lack physical transferability: they fit correlations across whole molecules or crystals rather than learning the quantum interactions between atomic pairs. Yet bonding, charge redistribution, orbital hybridization, and electronic coupling all emerge from these two-body interactions that define local quantum fields in many-body systems. We introduce QuantumCanvas, a large-scale multimodal benchmark that treats two-body quantum systems as foundational units of matter. The dataset spans 2,850 element-element pairs, each annotated with 18 electronic, thermodynamic, and geometric properties and paired with ten-channel image representations derived from l- and m-resolved orbital densities, angular field transforms, co-occupancy maps, and charge-density projections. These physically grounded images encode spatial, angular, and electrostatic symmetries without explicit coordinates, providing an interpretable visual modality for quantum learning. Benchmarking eight architectures across 18 targets, we report mean absolute errors of 0.201 eV on energy gap using GATv2, 0.265 eV on HOMO and 0.274 eV on LUMO using EGNN. For energy-related quantities, DimeNet attains 2.27 eV total-energy MAE and 0.132 eV repulsive-energy MAE, while a multimodal fusion model achieves a 2.15 eV Mermin free-energy MAE. Pretraining on QuantumCanvas further improves convergence stability and generalization when fine-tuned on larger datasets such as QM9, MD17, and CrysMTM. By unifying orbital physics with vision-based representation learning, QuantumCanvas provides a principled and interpretable basis for learning transferable quantum interactions through coupled visual and numerical modalities. Dataset and model implementations are available at https://github.com/KurbanIntelligenceLab/QuantumCanvas.

Paperid: 396, https://arxiv.org/pdf/2512.01512.pdf GitHub

Abstract:
Multimodal Large Language Models (MLLMs) have achieved great success in Speech-to-Text Translation (S2TT) tasks. However, current research is constrained by two key challenges: language coverage and efficiency. Most of the popular S2TT datasets are substantially English-centric, which restricts the scaling-up of MLLMs' many-to-many translation capabilities. Moreover, the inference speed of MLLMs degrades dramatically when the speech is converted into long sequences (e.g., 750 tokens). To address these limitations, we propose a Multilingual Cost-effective Accelerated Speech-to-Text Translator (MCAT) framework, which includes two innovations. First, a language scaling method that leverages curriculum learning and a data balancing strategy is introduced to extend the language coverage supported by MLLMs to 70 languages and achieve mutual translation among these languages. Second, an optimized speech adapter module is designed to reduce the length of the speech sequence to only 30 tokens. Extensive experiments were conducted on MLLMs of different scales (9B and 27B). The experimental results demonstrate that MCAT not only surpasses state-of-the-art end-to-end models on the FLEURS dataset across 70x69 directions but also enhances batch inference efficiency. This is achieved with only ~100M trainable parameters and by using only 10 hours of S2TT data per language. Furthermore, we have released MCAT as open-source to promote the development of MLLMs for robust S2TT capabilities. The code and models are released at https://github.com/yxduir/m2m-70.

Paperid: 397, https://arxiv.org/pdf/2512.01461.pdf GitHub

Abstract:
Model merging has emerged as a promising paradigm for enabling multi-task capabilities without additional training. However, existing methods often experience substantial performance degradation compared with individually fine-tuned models, even on similar tasks, underscoring the need to preserve task-specific information. This paper proposes Decomposition, Thresholding, and Scaling (DTS), an approximation-based personalized merging framework that preserves task-specific information with minimal storage overhead. DTS first applies singular value decomposition to the task-specific information and retains only a small subset of singular values and vectors. It then introduces a novel thresholding strategy that partitions singular vector elements into groups and assigns a scaling factor to each group. To enable generalization to unseen tasks, we further extend DTS with a variant that fuses task-specific information in a data-free manner based on the semantic similarity of task characteristics. Extensive experiments demonstrate that DTS consistently outperforms state-of-the-art baselines while requiring only 1\% additional storage per task. Furthermore, experiments on unseen tasks show that the DTS variant achieves significantly better generalization performance. Our code is available at https://github.com/krumpguo/DTS.

Paperid: 398, https://arxiv.org/pdf/2512.01443.pdf GitHub

Abstract:
We present Conformer-based decoders for the LibriBrain 2025 PNPL competition, targeting two foundational MEG tasks: Speech Detection and Phoneme Classification. Our approach adapts a compact Conformer to raw 306-channel MEG signals, with a lightweight convolutional projection layer and task-specific heads. For Speech Detection, a MEG-oriented SpecAugment provided a first exploration of MEG-specific augmentation. For Phoneme Classification, we used inverse-square-root class weighting and a dynamic grouping loader to handle 100-sample averaged examples. In addition, a simple instance-level normalization proved critical to mitigate distribution shifts on the holdout split. Using the official Standard track splits and F1-macro for model selection, our best systems achieved 88.9% (Speech) and 65.8% (Phoneme) on the leaderboard, surpassing the competition baselines and ranking within the top-10 in both tasks. For further implementation details, the technical documentation, source code, and checkpoints are available at https://github.com/neural2speech/libribrain-experiments.

Paperid: 399, https://arxiv.org/pdf/2512.01422.pdf GitHub

Abstract:
Mask Diffusion Models (MDMs) have recently emerged as a promising alternative to auto-regressive models (ARMs) for vision-language tasks, owing to their flexible balance of efficiency and accuracy. In this paper, for the first time, we introduce MDMs into the Scene Text Recognition (STR) task. We show that vanilla MDM lags behind ARMs in terms of accuracy, although it improves recognition efficiency. To bridge this gap, we propose MDiff4STR, a Mask Diffusion model enhanced with two key improvement strategies tailored for STR. Specifically, we identify two key challenges in applying MDMs to STR: noising gap between training and inference, and overconfident predictions during inference. Both significantly hinder the performance of MDMs. To mitigate the first issue, we develop six noising strategies that better align training with inference behavior. For the second, we propose a token-replacement noise mechanism that provides a non-mask noise type, encouraging the model to reconsider and revise overly confident but incorrect predictions. We conduct extensive evaluations of MDiff4STR on both standard and challenging STR benchmarks, covering diverse scenarios including irregular, artistic, occluded, and Chinese text, as well as whether the use of pretraining. Across these settings, MDiff4STR consistently outperforms popular STR models, surpassing state-of-the-art ARMs in accuracy, while maintaining fast inference with only three denoising steps. Code: https://github.com/Topdu/OpenOCR.

Paperid: 400, https://arxiv.org/pdf/2512.01390.pdf GitHub

Abstract:
Real-image super-resolution (Real-ISR) seeks to recover HR images from LR inputs with mixed, unknown degradations. While diffusion models surpass GANs in perceptual quality, they under-reconstruct high-frequency (HF) details due to a low-frequency (LF) bias and a depth-wise "low-first, high-later" hierarchy. We introduce FRAMER, a plug-and-play training scheme that exploits diffusion priors without changing the backbone or inference. At each denoising step, the final-layer feature map teaches all intermediate layers. Teacher and student feature maps are decomposed into LF/HF bands via FFT masks to align supervision with the model's internal frequency hierarchy. For LF, an Intra Contrastive Loss (IntraCL) stabilizes globally shared structure. For HF, an Inter Contrastive Loss (InterCL) sharpens instance-specific details using random-layer and in-batch negatives. Two adaptive modulators, Frequency-based Adaptive Weight (FAW) and Frequency-based Alignment Modulation (FAM), reweight per-layer LF/HF signals and gate distillation by current similarity. Across U-Net and DiT backbones (e.g., Stable Diffusion 2, 3), FRAMER consistently improves PSNR/SSIM and perceptual metrics (LPIPS, NIQE, MANIQA, MUSIQ). Ablations validate the final-layer teacher and random-layer negatives.

Paperid: 401, https://arxiv.org/pdf/2512.01353.pdf GitHub

Abstract:
Large language models (LLMs) remain vulnerable to jailbreak attacks that bypass safety guardrails to elicit harmful outputs. Existing approaches overwhelmingly operate within the prompt-optimization paradigm: whether through traditional algorithmic search or recent agent-based workflows, the resulting prompts typically retain malicious semantic signals that modern guardrails are primed to detect. In contrast, we identify a deeper, largely overlooked vulnerability stemming from the highly interconnected nature of an LLM's internal knowledge. This structure allows harmful objectives to be realized by weaving together sequences of benign sub-queries, each of which individually evades detection. To exploit this loophole, we introduce the Correlated Knowledge Attack Agent (CKA-Agent), a dynamic framework that reframes jailbreaking as an adaptive, tree-structured exploration of the target model's knowledge base. The CKA-Agent issues locally innocuous queries, uses model responses to guide exploration across multiple paths, and ultimately assembles the aggregated information to achieve the original harmful objective. Evaluated across state-of-the-art commercial LLMs (Gemini2.5-Flash/Pro, GPT-oss-120B, Claude-Haiku-4.5), CKA-Agent consistently achieves over 95% success rates even against strong guardrails, underscoring the severity of this vulnerability and the urgent need for defenses against such knowledge-decomposition attacks. Our codes are available at https://github.com/Graph-COM/CKA-Agent.

Paperid: 402, https://arxiv.org/pdf/2512.01329.pdf GitHub

Abstract:
Topology-consistent dynamic model sequences are essential for applications such as animation and model editing. However, existing 4D reconstruction methods face challenges in generating high-quality topology-consistent meshes. To address this, we propose a topology-aware dynamic reconstruction framework based on Gaussian Splatting. We introduce a Gaussian topological structure that explicitly encodes spatial connectivity. This structure enables topology-aware densification and pruning, preserving the manifold consistency of the Gaussian representation. Temporal regularization terms further ensure topological coherence over time, while differentiable mesh rasterization improves mesh quality. Experimental results demonstrate that our method reconstructs topology-consistent mesh sequences with significantly higher accuracy than existing approaches. Moreover, the resulting meshes enable precise 3D keypoint tracking. Project page: https://haza628.github.io/tagSplat/

Paperid: 403, https://arxiv.org/pdf/2512.01319.pdf GitHub

Abstract:
The precise segmentation of intracranial aneurysms and their parent vessels (IA-Vessel) is a critical step for hemodynamic analyses, which mainly depends on computational fluid dynamics (CFD). However, current segmentation methods predominantly focus on image-based evaluation metrics, often neglecting their practical effectiveness in subsequent CFD applications. To address this deficiency, we present the Intracranial Aneurysm Vessel Segmentation (IAVS) dataset, the first comprehensive, multi-center collection comprising 641 3D MRA images with 587 annotations of aneurysms and IA-Vessels. In addition to image-mask pairs, IAVS dataset includes detailed hemodynamic analysis outcomes, addressing the limitations of existing datasets that neglect topological integrity and CFD applicability. To facilitate the development and evaluation of clinically relevant techniques, we construct two evaluation benchmarks including global localization of aneurysms (Stage I) and fine-grained segmentation of IA-Vessel (Stage II) and develop a simple and effective two-stage framework, which can be used as a out-of-the-box method and strong baseline. For comprehensive evaluation of applicability of segmentation results, we establish a standardized CFD applicability evaluation system that enables the automated and consistent conversion of segmentation masks into CFD models, offering an applicability-focused assessment of segmentation outcomes. The dataset, code, and model will be public available at https://github.com/AbsoluteResonance/IAVS.

Paperid: 404, https://arxiv.org/pdf/2512.01311.pdf GitHub

Abstract:
Large language model based agents are increasingly deployed in complex, tool augmented environments. While reinforcement learning provides a principled mechanism for such agents to improve through interaction, its effectiveness critically depends on the availability of structured training tasks. In many realistic settings, however, no such tasks exist a challenge we term task scarcity, which has become a key bottleneck for scaling agentic RL. Existing approaches typically assume predefined task collections, an assumption that fails in novel environments where tool semantics and affordances are initially unknown. To address this limitation, we formalize the problem of Task Generation for Agentic RL, where an agent must learn within a given environment that lacks predefined tasks. We propose CuES, a Curiosity driven and Environment grounded Synthesis framework that autonomously generates diverse, executable, and meaningful tasks directly from the environment structure and affordances, without relying on handcrafted seeds or external corpora. CuES drives exploration through intrinsic curiosity, abstracts interaction patterns into reusable task schemas, and refines them through lightweight top down guidance and memory based quality control. Across three representative environments, AppWorld, BFCL, and WebShop, CuES produces task distributions that match or surpass manually curated datasets in both diversity and executability, yielding substantial downstream policy improvements. These results demonstrate that curiosity driven, environment grounded task generation provides a scalable foundation for agents that not only learn how to act, but also learn what to learn. The code is available at https://github.com/modelscope/AgentEvolver/tree/main/research/CuES.

Paperid: 405, https://arxiv.org/pdf/2512.01298.pdf GitHub

Abstract:
Temporal Action Localization (TAL) remains a fundamental challenge in video understanding, aiming to identify the start time, end time, and category of all action instances within untrimmed videos. While recent single-stage, anchor-free models like ActionFormer have set a high standard by leveraging Transformers for temporal reasoning, they often struggle with two persistent issues: the precise localization of actions with ambiguous or "fuzzy" temporal boundaries and the effective fusion of multi-scale contextual information. In this paper, we introduce the Temporal Boundary Transformer (TBT-Former), a new architecture that directly addresses these limitations. TBT-Former enhances the strong ActionFormer baseline with three core contributions: (1) a higher-capacity scaled Transformer backbone with an increased number of attention heads and an expanded Multi-Layer Perceptron (MLP) dimension for more powerful temporal feature extraction; (2) a cross-scale feature pyramid network (FPN) that integrates a top-down pathway with lateral connections, enabling richer fusion of high-level semantics and low-level temporal details; and (3) a novel boundary distribution regression head. Inspired by the principles of Generalized Focal Loss (GFL), this new head recasts the challenging task of boundary regression as a more flexible probability distribution learning problem, allowing the model to explicitly represent and reason about boundary uncertainty. Within the paradigm of Transformer-based architectures, TBT-Former advances the formidable benchmark set by its predecessors, establishing a new level of performance on the highly competitive THUMOS14 and EPIC-Kitchens 100 datasets, while remaining competitive on the large-scale ActivityNet-1.3. Our code is available at https://github.com/aaivu/In21-S7-CS4681-AML-Research-Projects/tree/main/projects/210536K-Multi-Modal-Learning_Video-Understanding

Paperid: 406, https://arxiv.org/pdf/2512.01296.pdf GitHub

Abstract:
Real-time 3D reconstruction is a fundamental task in computer graphics. Recently, differentiable-rendering-based SLAM system has demonstrated significant potential, enabling photorealistic scene rendering through learnable scene representations such as Neural Radiance Fields (NeRF) and 3D Gaussian Splatting (3DGS). Current differentiable rendering methods face dual challenges in real-time computation and sensor noise sensitivity, leading to degraded geometric fidelity in scene reconstruction and limited practicality. To address these challenges, we propose a novel real-time system EGG-Fusion, featuring robust sparse-to-dense camera tracking and a geometry-aware Gaussian surfel mapping module, introducing an information filter-based fusion method that explicitly accounts for sensor noise to achieve high-precision surface reconstruction. The proposed differentiable Gaussian surfel mapping effectively models multi-view consistent surfaces while enabling efficient parameter optimization. Extensive experimental results demonstrate that the proposed system achieves a surface reconstruction error of 0.6\textit{cm} on standardized benchmark datasets including Replica and ScanNet++, representing over 20\% improvement in accuracy compared to state-of-the-art (SOTA) GS-based methods. Notably, the system maintains real-time processing capabilities at 24 FPS, establishing it as one of the most accurate differentiable-rendering-based real-time reconstruction systems. Project Page: https://zju3dv.github.io/eggfusion/

Paperid: 407, https://arxiv.org/pdf/2512.01282.pdf GitHub

Abstract:
As web platforms evolve towards greater personalization and emotional complexity, conversational agents must transcend superficial empathy to demonstrate identity-aware emotional reasoning. However, existing systems face two limitations: (1) reliance on situation-centric datasets lacking persistent user identity, which hampers the capture of personalized affective nuances; and (2) dependence on opaque, coarse reward signals that hinder development of verifiable empathetic reasoning. To address these gaps, we introduce KardiaBench, a large-scale user-grounded benchmark comprising 178,080 QA pairs across 22,080 multi-turn conversations anchored to 671 real-world profiles. The dataset is constructed via a model-in-the-loop pipeline with iterative rubric-guided refinement to ensure psychological plausibility and persona consistency. This progressive empathy pipeline that integrates user comprehension, contextual reasoning, and emotion perception into conversations, followed by iterative critique and rubric-based refinement to ensure psychological plausibility, emotional fidelity, and persona consistency. Building on this, we propose Kardia-R1, a framework that trains models for interpretable, stepwise empathetic cognition. Kardia-R1 leverages Rubric-as-Judge Empathetic Reinforcement Learning (Rubric-ERL), a GRPO-based method that uses explainable, human-aligned rubric rewards to tightly couple user understanding, emotional inference, and supportive response generation. Extensive experiments across four LLM backbones demonstrate that Kardia-R1 consistently outperforms othet methods in emotion accuracy, empathy, relevance, persona consistency, and safety. Our dataset and model will be released at https://github.com/JhCircle/Kardia-R1.

Paperid: 408, https://arxiv.org/pdf/2512.01252.pdf GitHub

Abstract:
Recent efforts on Diffusion Mixture-of-Experts (MoE) models have primarily focused on developing more sophisticated routing mechanisms. However, we observe that the underlying architectural configuration space remains markedly under-explored. Inspired by the MoE design paradigms established in large language models (LLMs), we identify a set of crucial architectural factors for building effective Diffusion MoE models--including DeepSeek-style expert modules, alternative intermediate widths, varying expert counts, and enhanced attention positional encodings. Our systematic study reveals that carefully tuning these configurations is essential for unlocking the full potential of Diffusion MoE models, often yielding gains that exceed those achieved by routing innovations alone. Through extensive experiments, we present novel architectures that can be efficiently applied to both latent and pixel-space diffusion frameworks, which provide a practical and efficient training recipe that enables Diffusion MoE models to surpass strong baselines while using equal or fewer activated parameters. All code and models are publicly available at: https://github.com/yhlleo/EfficientMoE.

Paperid: 409, https://arxiv.org/pdf/2512.01248.pdf GitHub

Abstract:
Table recognition (TR) aims to transform table images into semi-structured representations such as HTML or Markdown. As a core component of document parsing, TR has long relied on supervised learning, with recent efforts dominated by fine-tuning vision-language models (VLMs) using labeled data. While VLMs have brought TR to the next level, pushing performance further demands large-scale labeled data that is costly to obtain. Consequently, although proprietary models have continuously pushed the performance boundary, open-source models, often trained with limited resources and, in practice, the only viable option for many due to privacy regulations, still lag far behind. To bridge this gap, we introduce TRivia, a self-supervised fine-tuning method that enables pretrained VLMs to learn TR directly from unlabeled table images in the wild. Built upon Group Relative Policy Optimization, TRivia automatically identifies unlabeled samples that most effectively facilitate learning and eliminates the need for human annotations through a question-answering-based reward mechanism. An attention-guided module generates diverse questions for each table image, and the ability to interpret the recognition results and answer them correctly provides feedback to optimize the TR model. This closed-loop process allows the TR model to autonomously learn to recognize, structure, and reason over tables without labeled data. Leveraging this pipeline, we present TRivia-3B, an open-sourced, compact, and state-of-the-art TR model that surpasses existing systems (e.g., Gemini 2.5 Pro, MinerU2.5) on three popular benchmarks. Model and code are released at: https://github.com/opendatalab/TRivia

Paperid: 410, https://arxiv.org/pdf/2512.01236.pdf GitHub

Abstract:
Personalized generation models for a single subject have demonstrated remarkable effectiveness, highlighting their significant potential. However, when extended to multiple subjects, existing models often exhibit degraded performance, particularly in maintaining subject consistency and adhering to textual prompts. We attribute these limitations to the absence of high-quality multi-subject datasets and refined post-training strategies. To address these challenges, we propose a scalable multi-subject data generation pipeline that leverages powerful single-subject generation models to construct diverse and high-quality multi-subject training data. Through this dataset, we first enable single-subject personalization models to acquire knowledge of synthesizing multi-image and multi-subject scenarios. Furthermore, to enhance both subject consistency and text controllability, we design a set of Pairwise Subject-Consistency Rewards and general-purpose rewards, which are incorporated into a refined reinforcement learning stage. To comprehensively evaluate multi-subject personalization, we introduce a new benchmark that assesses model performance using seven subsets across three dimensions. Extensive experiments demonstrate the effectiveness of our approach in advancing multi-subject personalized image generation. Github Link: https://github.com/wang-shulei/PSR

Paperid: 411, https://arxiv.org/pdf/2512.01204.pdf GitHub

Abstract:
Generating high-fidelity, physically interactive 3D simulated tabletop scenes is essential for embodied AI -- especially for robotic manipulation policy learning and data synthesis. However, current text- or image-driven 3D scene generation methods mainly focus on large-scale scenes, struggling to capture the high-density layouts and complex spatial relations that characterize tabletop scenes. To address these challenges, we propose TabletopGen, a training-free, fully automatic framework that generates diverse, instance-level interactive 3D tabletop scenes. TabletopGen accepts a reference image as input, which can be synthesized by a text-to-image model to enhance scene diversity. We then perform instance segmentation and completion on the reference to obtain per-instance images. Each instance is reconstructed into a 3D model followed by canonical coordinate alignment. The aligned 3D models then undergo pose and scale estimation before being assembled into a collision-free, simulation-ready tabletop scene. A key component of our framework is a novel pose and scale alignment approach that decouples the complex spatial reasoning into two stages: a Differentiable Rotation Optimizer for precise rotation recovery and a Top-view Spatial Alignment mechanism for robust translation and scale estimation, enabling accurate 3D reconstruction from 2D reference. Extensive experiments and user studies show that TabletopGen achieves state-of-the-art performance, markedly surpassing existing methods in visual fidelity, layout accuracy, and physical plausibility, capable of generating realistic tabletop scenes with rich stylistic and spatial diversity. Our code will be publicly available.

Paperid: 412, https://arxiv.org/pdf/2512.01199.pdf GitHub

Abstract:
Neurons process information in ways that depend on their cell type, connectivity, and the brain region in which they are embedded. However, inferring these factors from neural activity remains a significant challenge. To build general-purpose representations that allow for resolving information about a neuron's identity, we introduce NuCLR, a self-supervised framework that aims to learn representations of neural activity that allow for differentiating one neuron from the rest. NuCLR brings together views of the same neuron observed at different times and across different stimuli and uses a contrastive objective to pull these representations together. To capture population context without assuming any fixed neuron ordering, we build a spatiotemporal transformer that integrates activity in a permutation-equivariant manner. Across multiple electrophysiology and calcium imaging datasets, a linear decoding evaluation on top of NuCLR representations achieves a new state-of-the-art for both cell type and brain region decoding tasks, and demonstrates strong zero-shot generalization to unseen animals. We present the first systematic scaling analysis for neuron-level representation learning, showing that increasing the number of animals used during pretraining consistently improves downstream performance. The learned representations are also label-efficient, requiring only a small fraction of labeled samples to achieve competitive performance. These results highlight how large, diverse neural datasets enable models to recover information about neuron identity that generalize across animals. Code is available at https://github.com/nerdslab/nuclr.

Paperid: 413, https://arxiv.org/pdf/2512.01188.pdf GitHub

Abstract:
A robot's instantaneous sensory observations do not always reveal task-relevant state information. Under such partial observability, optimal behavior typically involves explicitly acting to gain the missing information. Today's standard robot learning techniques struggle to produce such active perception behaviors. We propose a simple real-world robot learning recipe to efficiently train active perception policies. Our approach, asymmetric advantage weighted regression (AAWR), exploits access to "privileged" extra sensors at training time. The privileged sensors enable training high-quality privileged value functions that aid in estimating the advantage of the target policy. Bootstrapping from a small number of potentially suboptimal demonstrations and an easy-to-obtain coarse policy initialization, AAWR quickly acquires active perception behaviors and boosts task performance. In evaluations on 8 manipulation tasks on 3 robots spanning varying degrees of partial observability, AAWR synthesizes reliable active perception behaviors that outperform all prior approaches. When initialized with a "generalist" robot policy that struggles with active perception tasks, AAWR efficiently generates information-gathering behaviors that allow it to operate under severe partial observability for manipulation tasks. Website: https://penn-pal-lab.github.io/aawr/

Paperid: 414, https://arxiv.org/pdf/2512.01179.pdf GitHub

Abstract:
This research designs a unified architecture of CTR prediction benchmark (Bench-CTR) platform that offers flexible interfaces with datasets and components of a wide range of CTR prediction models. Moreover, we construct a comprehensive system of evaluation protocols encompassing real-world and synthetic datasets, a taxonomy of metrics, standardized procedures and experimental guidelines for calibrating the performance of CTR prediction models. Furthermore, we implement the proposed benchmark platform and conduct a comparative study to evaluate a wide range of state-of-the-art models from traditional multivariate statistical to modern large language model (LLM)-based approaches on three public datasets and two synthetic datasets. Experimental results reveal that, (1) high-order models largely outperform low-order models, though such advantage varies in terms of metrics and on different datasets; (2) LLM-based models demonstrate a remarkable data efficiency, i.e., achieving the comparable performance to other models while using only 2% of the training data; (3) the performance of CTR prediction models has achieved significant improvements from 2015 to 2016, then reached a stage with slow progress, which is consistent across various datasets. This benchmark is expected to facilitate model development and evaluation and enhance practitioners' understanding of the underlying mechanisms of models in the area of CTR prediction. Code is available at https://github.com/NuriaNinja/Bench-CTR.

Paperid: 415, https://arxiv.org/pdf/2512.01178.pdf GitHub

Abstract:
Monocular 3D object detection is a fundamental yet challenging task in 3D scene understanding. Existing approaches heavily depend on supervised learning with extensive 3D annotations, which are often acquired from LiDAR point clouds through labor-intensive labeling processes. To tackle this problem, we propose VSRD++, a novel weakly supervised framework for monocular 3D object detection that eliminates the reliance on 3D annotations and leverages neural-field-based volumetric rendering with weak 2D supervision. VSRD++ consists of a two-stage pipeline: multi-view 3D autolabeling and subsequent monocular 3D detector training. In the multi-view autolabeling stage, object surfaces are represented as signed distance fields (SDFs) and rendered as instance masks via the proposed instance-aware volumetric silhouette rendering. To optimize 3D bounding boxes, we decompose each instance's SDF into a cuboid SDF and a residual distance field (RDF) that captures deviations from the cuboid. To address the geometry inconsistency commonly observed in volume rendering methods applied to dynamic objects, we model the dynamic objects by including velocity into bounding box attributes as well as assigning confidence to each pseudo-label. Moreover, we also employ a 3D attribute initialization module to initialize the dynamic bounding box parameters. In the monocular 3D object detection phase, the optimized 3D bounding boxes serve as pseudo labels for training monocular 3D object detectors. Extensive experiments on the KITTI-360 dataset demonstrate that VSRD++ significantly outperforms existing weakly supervised approaches for monocular 3D object detection on both static and dynamic scenes. Code is available at https://github.com/Magicboomliu/VSRD_plus_plus

Paperid: 416, https://arxiv.org/pdf/2512.01174.pdf GitHub

Abstract:
As agentic AI systems increasingly operate autonomously, establishing trust through verifiable evaluation becomes critical. Yet existing benchmarks lack the transparency and auditability needed to assess whether agents behave reliably. We present DrawingBench, a verification framework for evaluating the trustworthiness of agentic LLMs through spatial reasoning tasks that require generating sequences of low-level GUI actions. Unlike opaque evaluations, DrawingBench provides transparent, rule-based assessment: 8 objective criteria enable reproducible scoring, while action-level inspection allows stakeholders to audit agent behavior. Our framework comprises 250 diverse prompts across 20 categories and 4 difficulty levels, deterministic evaluation metrics, and an external oversight mechanism through multi-turn feedback that enables human control over agent refinement. Evaluating four state-of-the-art LLMs (Claude-4 Sonnet, GPT-4.1, GPT-4.1-mini, Gemini-2.5 Flash) across 1,000 tests, we establish both capabilities and limitations: models achieved 92.8% perfect performance with structured external feedback driving significant improvements (average +3.2%, up to +32.8% for complex scenes), but systematic error patterns emerged in tool state management and long-horizon planning. Notably, specification clarity proved more important than task complexity -- models achieved 100% perfect performance when given explicit, verifiable criteria. These findings demonstrate that transparent evaluation frameworks can establish trust in agentic systems, with external oversight proving more reliable than self-correction for guiding agent behavior. Our open-source framework provides a template for trustworthy agent assessment. Code and data: https://github.com/hyunjun1121/DrawingBench

Paperid: 417, https://arxiv.org/pdf/2512.01128.pdf GitHub

Abstract:
Face forgery detection encompasses multiple critical tasks, including identifying forged images and videos and localizing manipulated regions and temporal segments. Current approaches typically employ task-specific models with independent architectures, leading to computational redundancy and ignoring potential correlations across related tasks. We introduce OmniFD, a unified framework that jointly addresses four core face forgery detection tasks within a single model, i.e., image and video classification, spatial localization, and temporal localization. Our architecture consists of three principal components: (1) a shared Swin Transformer encoder that extracts unified 4D spatiotemporal representations from both images and video inputs, (2) a cross-task interaction module with learnable queries that dynamically captures inter-task dependencies through attention-based reasoning, and (3) lightweight decoding heads that transform refined representations into corresponding predictions for all FFD tasks. Extensive experiments demonstrate OmniFD's advantage over task-specific models. Its unified design leverages multi-task learning to capture generalized representations across tasks, especially enabling fine-grained knowledge transfer that facilitates other tasks. For example, video classification accuracy improves by 4.63% when image data are incorporated. Furthermore, by unifying images, videos and the four tasks within one framework, OmniFD achieves superior performance across diverse benchmarks with high efficiency and scalability, e.g., reducing 63% model parameters and 50% training time. It establishes a practical and generalizable solution for comprehensive face forgery detection in real-world applications. The source code is made available at https://github.com/haotianll/OmniFD.

Paperid: 418, https://arxiv.org/pdf/2512.01119.pdf GitHub

Abstract:
AI systems deployed in the real world must contend with distractions and out-of-distribution (OOD) noise that can destabilize their policies and lead to unsafe behavior. While robust training can reduce sensitivity to some forms of noise, it is infeasible to anticipate all possible OOD conditions. To mitigate this issue, we develop an algorithm that leverages a world model's inherent measure of surprise to reduce the impact of noise in world model--based reinforcement learning agents. We introduce both multi-representation and single-representation rejection sampling, enabling robustness to settings with multiple faulty sensors or a single faulty sensor. While the introduction of noise typically degrades agent performance, we show that our techniques preserve performance relative to baselines under varying types and levels of noise across multiple environments within self-driving simulation domains (CARLA and Safety Gymnasium). Furthermore, we demonstrate that our methods enhance the stability of two state-of-the-art world models with markedly different underlying architectures: Cosmos and DreamerV3. Together, these results highlight the robustness of our approach across world modeling domains. We release our code at https://github.com/Bluefin-Tuna/WISER .

Paperid: 419, https://arxiv.org/pdf/2512.01109.pdf GitHub

Abstract:
In this work, we aim to clarify and reconcile metrics for evaluating privacy protection in text through a systematic survey. Although text anonymization is essential for enabling NLP research and model development in domains with sensitive data, evaluating whether anonymization methods sufficiently protect privacy remains an open challenge. In manually reviewing 47 papers that report privacy metrics, we identify and compare six distinct privacy notions, and analyze how the associated metrics capture different aspects of privacy risk. We then assess how well these notions align with legal privacy standards (HIPAA and GDPR), as well as user-centered expectations grounded in HCI studies. Our analysis offers practical guidance on navigating the landscape of privacy evaluation approaches further and highlights gaps in current practices. Ultimately, we aim to facilitate more robust, comparable, and legally aware privacy evaluations in text anonymization.

Paperid: 420, https://arxiv.org/pdf/2512.01059.pdf GitHub GitHub

Abstract:
Although scaling laws and many empirical results suggest that increasing the size of Vision Transformers often improves performance, model accuracy and training behavior are not always monotonically increasing with scale. Focusing on ViT-B/16 trained on ImageNet-1K, we study two simple parameter-reduction strategies applied to the MLP blocks, each removing 32.7\% of the baseline parameters. Our \emph{GroupedMLP} variant shares MLP weights between adjacent transformer blocks and achieves 81.47\% top-1 accuracy while maintaining the baseline computational cost. Our \emph{ShallowMLP} variant halves the MLP hidden dimension and reaches 81.25\% top-1 accuracy with a 38\% increase in inference throughput. Both models outperform the 86.6M-parameter baseline (81.05\%) and exhibit substantially improved training stability, reducing peak-to-final accuracy degradation from 0.47\% to the range 0.03\% to 0.06\%. These results suggest that, for ViT-B/16 on ImageNet-1K with a standard training recipe, the model operates in an overparameterized regime in which MLP capacity can be reduced without harming performance and can even slightly improve it. More broadly, our findings suggest that architectural constraints such as parameter sharing and reduced width may act as useful inductive biases, and highlight the importance of how parameters are allocated when designing Vision Transformers. All code is available at: https://github.com/AnanthaPadmanaban-KrishnaKumar/parameter-efficient-vit-mlps.

Paperid: 421, https://arxiv.org/pdf/2512.01048.pdf GitHub

Abstract:
Vision-language models (VLMs) have made great strides in addressing temporal understanding tasks, which involve characterizing visual changes across a sequence of images. However, recent works have suggested that when making predictions, VLMs may rely on static feature biases, such as background or object features, rather than dynamic visual changes. Static feature biases are a type of shortcut and can contribute to systematic prediction errors on downstream tasks; as a result, identifying and characterizing error-inducing static feature biases is critical prior to real-world model deployment. In this work, we introduce TRoVe, an automated approach for discovering error-inducing static feature biases learned by temporal VLMs. Given a trained VLM and an annotated validation dataset associated with a downstream classification task, TRoVe extracts candidate static features from the dataset and scores each feature by (i) the effect of the feature on classification errors as well as (ii) the extent to which the VLM relies on the feature when making predictions. In order to quantitatively evaluate TRoVe, we introduce an evaluation framework consisting of 101 trained temporal VLMs paired with ground-truth annotations for learned static feature biases. We use this framework to demonstrate that TRoVe can accurately identify error-inducing static feature biases in VLMs, achieving a 28.6% improvement over the closest baseline. Finally, we apply TRoVe to 7 off-the-shelf VLMs and 2 temporal understanding tasks, surfacing previously-unknown static feature biases and demonstrating that knowledge of learned biases can aid in improving model performance at test time. Our code is available at https://github.com/Stanford-AIMI/TRoVe.

Paperid: 422, https://arxiv.org/pdf/2512.01038.pdf GitHub

Abstract:
Foundation models (FMs) have opened new avenues for machine learning applications due to their ability to adapt to new and unseen tasks with minimal or no further training. Time-series foundation models (TSFMs) -- FMs trained on time-series data -- have shown strong performance on classification, regression, and imputation tasks. Recent pipelines combine TSFMs with task-specific encoders, decoders, and adapters to improve performance; however, assembling such pipelines typically requires ad hoc, model-specific implementations that hinder modularity and reproducibility. We introduce FMTK, an open-source, lightweight and extensible toolkit for constructing and fine-tuning TSFM pipelines via standardized backbone and component abstractions. FMTK enables flexible composition across models and tasks, achieving correctness and performance with an average of seven lines of code. https://github.com/umassos/FMTK

Paperid: 423, https://arxiv.org/pdf/2512.01031.pdf GitHub

Abstract:
Vision-Language-Action models (VLAs) are becoming increasingly capable across diverse robotic tasks. However, their real-world deployment remains slow and inefficient: demonstration videos are often sped up by 5-10x to appear smooth, with noticeable action stalls and delayed reactions to environmental changes. Asynchronous inference offers a promising solution to achieve continuous and low-latency control by enabling robots to execute actions and perform inference simultaneously. However, because the robot and environment continue to evolve during inference, a temporal misalignment arises between the prediction and execution intervals. This leads to significant action instability, while existing methods either degrade accuracy or introduce runtime overhead to mitigate it. We propose VLASH, a general asynchronous inference framework for VLAs that delivers smooth, accurate, and fast reaction control without additional overhead or architectural changes. VLASH estimates the future execution-time state by rolling the robot state forward with the previously generated action chunk, thereby bridging the gap between prediction and execution. Experiments show that VLASH achieves up to 2.03x speedup and reduces reaction latency by up to 17.4x compared to synchronous inference while fully preserving the original accuracy. Moreover, it empowers VLAs to handle fast-reaction, high-precision tasks such as playing ping-pong and playing whack-a-mole, where traditional synchronous inference fails. Code is available at https://github.com/mit-han-lab/vlash

Paperid: 424, https://arxiv.org/pdf/2512.01022.pdf GitHub

Abstract:
In this paper, we explore an important yet underexplored task in robot manipulation: cycle-based manipulation, where robots need to perform cyclic or repetitive actions with an expected terminal time. These tasks are crucial in daily life, such as shaking a bottle or knocking a nail. However, few prior works have explored this task, leading to two main challenges: 1) the imitation methods often fail to complete these tasks within the expected terminal time due to the ineffective utilization of history; 2) the absence of a benchmark with sufficient data and automatic evaluation tools hinders development of effective solutions in this area. To address these challenges, we first propose the CycleManip framework to achieve cycle-based task manipulation in an end-to-end imitation manner without requiring any extra models, hierarchical structure or significant computational overhead. The core insight is to enhance effective history perception by a cost-aware sampling strategy and to improve historical understanding by multi-task learning. Second, we introduce a cycle-based task manipulation benchmark, which provides diverse cycle-based tasks, and an automatic evaluation method. Extensive experiments conducted in both simulation and real-world settings demonstrate that our method achieves high success rates in cycle-based task manipulation. The results further show strong adaptability performance in general manipulation, and the plug-and-play ability on imitation policies such as Vision-Language-Action (VLA) models. Moreover, the results show that our approach can be applied across diverse robotic platforms, including bi-arm grippers, dexterous hands, and humanoid robots.

Paperid: 425, https://arxiv.org/pdf/2512.01008.pdf GitHub

Abstract:
Text-driven 3D reconstruction demands a mask generator that simultaneously understands open-vocabulary instructions and remains consistent across viewpoints. We present LISA-3D, a two-stage framework that lifts language-image segmentation into 3D by retrofitting the instruction-following model LISA with geometry-aware Low-Rank Adaptation (LoRA) layers and reusing a frozen SAM-3D reconstructor. During training we exploit off-the-shelf RGB-D sequences and their camera poses to build a differentiable reprojection loss that enforces cross-view agreement without requiring any additional 3D-text supervision. The resulting masks are concatenated with RGB images to form RGBA prompts for SAM-3D, which outputs Gaussian splats or textured meshes without retraining. Across ScanRefer and Nr3D, LISA-3D improves language-to-3D accuracy by up to +15.6 points over single-view baselines while adapting only 11.6M parameters. The system is modular, data-efficient, and supports zero-shot deployment on unseen categories, providing a practical recipe for language-guided 3D content creation. Our code will be available at https://github.com/binisalegend/LISA-3D.

Paperid: 426, https://arxiv.org/pdf/2512.00997.pdf GitHub

Abstract:
We introduce IndiMathBench, a human-verified benchmark designed to evaluate mathematical theorem proving, curated using an AI-powered human-assisted pipeline for formalizing natural language problems in Lean. IndiMathBench is composed of 312 formal Lean 4 theorems paired with their corresponding informal problem statements, sourced from Indian Mathematics Olympiads. Through category-based retrieval, iterative compiler feedback, and multi-model ensembles, our pipeline generates candidate formalizations that experts efficiently validate via an interactive dashboard with automated quality summaries. Evaluation across multiple frontier models demonstrates that autoformalization remains challenging, with substantial gaps between syntactic validity and semantic correctness, while theorem proving success rates remain low even with iterative refinement, demonstrating that \benchmark~presents a challenging testbed for mathematical reasoning. IndiMathBench is available at https://github.com/prmbiy/IndiMathBench.

Paperid: 427, https://arxiv.org/pdf/2512.00993.pdf GitHub

Abstract:
Composition matters during the photo-taking process, yet many casual users struggle to frame well-composed images. To provide composition guidance, we introduce PhotoFramer, a multi-modal composition instruction framework. Given a poorly composed image, PhotoFramer first describes how to improve the composition in natural language and then generates a well-composed example image. To train such a model, we curate a large-scale dataset. Inspired by how humans take photos, we organize composition guidance into a hierarchy of sub-tasks: shift, zoom-in, and view-change tasks. Shift and zoom-in data are sampled from existing cropping datasets, while view-change data are obtained via a two-stage pipeline. First, we sample pairs with varying viewpoints from multi-view datasets, and train a degradation model to transform well-composed photos into poorly composed ones. Second, we apply this degradation model to expert-taken photos to synthesize poor images to form training pairs. Using this dataset, we finetune a model that jointly processes and generates both text and images, enabling actionable textual guidance with illustrative examples. Extensive experiments demonstrate that textual instructions effectively steer image composition, and coupling them with exemplars yields consistent improvements over exemplar-only baselines. PhotoFramer offers a practical step toward composition assistants that make expert photographic priors accessible to everyday users. Codes, model weights, and datasets have been released in https://zhiyuanyou.github.io/photoframer.

Paperid: 428, https://arxiv.org/pdf/2512.00975.pdf GitHub

Abstract:
A generalist robotic policy needs both semantic understanding for task planning and the ability to interact with the environment through predictive capabilities. To tackle this, we present MM-ACT, a unified Vision-Language-Action (VLA) model that integrates text, image, and action in shared token space and performs generation across all three modalities. MM-ACT adopts a re-mask parallel decoding strategy for text and image generation, and employs a one-step parallel decoding strategy for action generation to improve efficiency. We introduce Context-Shared Multimodal Learning, a unified training paradigm that supervises generation in all three modalities from a shared context, enhancing action generation through cross-modal learning. Experiments were conducted on the LIBERO simulation and Franka real-robot setups as well as RoboTwin2.0 to assess in-domain and out-of-domain performances respectively. Our approach achieves a success rate of 96.3% on LIBERO, 72.0% across three tasks of real Franka, and 52.38% across eight bimanual tasks of RoboTwin2.0 with an additional gain of 9.25% from cross-modal learning. We release our codes, models and data at https://github.com/HHYHRHY/MM-ACT.

Paperid: 429, https://arxiv.org/pdf/2512.00974.pdf GitHub

Abstract:
Soft-core processors on resource-constrained FPGAs often suffer from low code density and reliance on proprietary toolchains. This paper details the design, implementation, and evaluation of a 32-bit dual-stack microprocessor architecture optimized for low-cost, resource-constrained Field-Programmable Gate Arrays (FPGAs). Implemented on the Gowin GW1NR-9 (Tang Nano 9K), the processor utilizes an instruction set architecture (ISA) inspired from a subset of the WebAssembly (WASM) specification to achieve high code density. Unlike traditional soft-cores that often rely on proprietary vendor toolchains and opaque IP blocks, this design is synthesized and routed utilizing an open-source flow, providing transparency and portability. The architecture features a dual-stack model (Data and Return), executing directly from SPI Flash via an Execute-in-Place (XIP) mechanism to conserve scarce Block RAM on the intended target device. An analysis of the trade-offs involved in stack depth parametrization is presented, demonstrating that an 8-entry distributed RAM implementation provides a balance between logic resource utilization ($\sim 80\%$) and routing congestion. Furthermore, timing hazards in single-cycle stack operations are identified and resolved through a refined Finite State Machine (FSM) design. The system achieves a stable operating frequency of 27 MHz, limited by Flash latency, and successfully executes simple applications including a single and multi-digit infix calculator.

Paperid: 430, https://arxiv.org/pdf/2512.00960.pdf GitHub

Abstract:
Generalized robots must learn from diverse, large-scale human-object interactions (HOI) to operate robustly in the real world. Monocular internet videos offer a nearly limitless and readily available source of data, capturing an unparalleled diversity of human activities, objects, and environments. However, accurately and scalably extracting 4D interaction data from these in-the-wild videos remains a significant and unsolved challenge. Thus, in this work, we introduce 4DHOISolver, a novel and efficient optimization framework that constrains the ill-posed 4D HOI reconstruction problem by leveraging sparse, human-in-the-loop contact point annotations, while maintaining high spatio-temporal coherence and physical plausibility. Leveraging this framework, we introduce Open4DHOI, a new large-scale 4D HOI dataset featuring a diverse catalog of 144 object types and 103 actions. Furthermore, we demonstrate the effectiveness of our reconstructions by enabling an RL-based agent to imitate the recovered motions. However, a comprehensive benchmark of existing 3D foundation models indicates that automatically predicting precise human-object contact correspondences remains an unsolved problem, underscoring the immediate necessity of our human-in-the-loop strategy while posing an open challenge to the community. Data and code will be publicly available at https://wenboran2002.github.io/open4dhoi/

Paperid: 431, https://arxiv.org/pdf/2512.00953.pdf GitHub

Abstract:
In the domain of moment retrieval, accurately identifying temporal segments within videos based on natural language queries remains challenging. Traditional methods often employ pre-trained models that struggle with fine-grained information and deterministic reasoning, leading to difficulties in aligning with complex or ambiguous moments. To overcome these limitations, we explore Deep Evidential Regression (DER) to construct a vanilla Evidential baseline. However, this approach encounters two major issues: the inability to effectively handle modality imbalance and the structural differences in DER's heuristic uncertainty regularizer, which adversely affect uncertainty estimation. This misalignment results in high uncertainty being incorrectly associated with accurate samples rather than challenging ones. Our observations indicate that existing methods lack the adaptability required for complex video scenarios. In response, we propose Debiased Evidential Learning for Moment Retrieval (DEMR), a novel framework that incorporates a Reflective Flipped Fusion (RFF) block for cross-modal alignment and a query reconstruction task to enhance text sensitivity, thereby reducing bias in uncertainty estimation. Additionally, we introduce a Geom-regularizer to refine uncertainty predictions, enabling adaptive alignment with difficult moments and improving retrieval accuracy. Extensive testing on standard datasets and debiased datasets ActivityNet-CD and Charades-CD demonstrates significant enhancements in effectiveness, robustness, and interpretability, positioning our approach as a promising solution for temporal-semantic robustness in moment retrieval. The code is publicly available at https://github.com/KaijingOfficial/DEMR.

Paperid: 432, https://arxiv.org/pdf/2512.00936.pdf GitHub

Abstract:
Grounding complex, compositional visual queries with multiple objects and relationships is a fundamental challenge for vision-language models. While standard phrase grounding methods excel at localizing single objects, they lack the structural inductive bias to parse intricate relational descriptions, often failing as queries become more descriptive. To address this structural deficit, we focus on scene-graph grounding, a powerful but less-explored formulation where the query is an explicit graph of objects and their relationships. However, existing methods for this task also struggle, paradoxically showing decreased performance as the query graph grows -- failing to leverage the very information that should make grounding easier. We introduce SceneProp, a novel method that resolves this issue by reformulating scene-graph grounding as a Maximum a Posteriori (MAP) inference problem in a Markov Random Field (MRF). By performing global inference over the entire query graph, SceneProp finds the optimal assignment of image regions to nodes that jointly satisfies all constraints. This is achieved within an end-to-end framework via a differentiable implementation of the Belief Propagation algorithm. Experiments on four benchmarks show that our dedicated focus on the scene-graph grounding formulation allows SceneProp to significantly outperform prior work. Critically, its accuracy consistently improves with the size and complexity of the query graph, demonstrating for the first time that more relational context can, and should, lead to better grounding. Codes are available at https://github.com/keitaotani/SceneProp.

Paperid: 433, https://arxiv.org/pdf/2512.00911.pdf GitHub

Abstract:
Panoramic cameras, capable of capturing a 360-degree field of view, are crucial in robotic vision, particularly in environments with sparse features. However, non-upright panoramas due to unstable robot postures hinder downstream tasks. Traditional IMU-based correction methods suffer from drift and external disturbances, while vision-based approaches offer a promising alternative. This study presents a dual-stream angle-aware generation network that jointly estimates camera inclination angles and reconstructs upright panoramic images. The network comprises a CNN branch that extracts local geometric structures from equirectangular projections and a ViT branch that captures global contextual cues from cubemap projections. These are integrated through a dual-projection adaptive fusion module that aligns spatial features across both domains. To further enhance performance, we introduce a high-frequency enhancement block, circular padding, and channel attention mechanisms to preserve 360° continuity and improve geometric sensitivity. Experiments on the SUN360 and M3D datasets demonstrate that our method outperforms existing approaches in both inclination estimation and upright panorama generation. Ablation studies further validate the contribution of each module and highlight the synergy between the two tasks. The code and related datasets can be found at: https://github.com/YuhaoShine/DualProjectionFusion.

Paperid: 434, https://arxiv.org/pdf/2512.00909.pdf GitHub

Abstract:
Recent advancements in diffusion models have significantly improved the realism and generalizability of character-driven animation, enabling the synthesis of high-quality motion from just a single RGB image and a set of driving poses. Nevertheless, generating temporally coherent long-form content remains challenging. Existing approaches are constrained by computational and memory limitations, as they are typically trained on short video segments, thus performing effectively only over limited frame lengths and hindering their potential for extended coherent generation. To address these constraints, we propose TalkingPose, a novel diffusion-based framework specifically designed for producing long-form, temporally consistent human upper-body animations. TalkingPose leverages driving frames to precisely capture expressive facial and hand movements, transferring these seamlessly to a target actor through a stable diffusion backbone. To ensure continuous motion and enhance temporal coherence, we introduce a feedback-driven mechanism built upon image-based diffusion models. Notably, this mechanism does not incur additional computational costs or require secondary training stages, enabling the generation of animations with unlimited duration. Additionally, we introduce a comprehensive, large-scale dataset to serve as a new benchmark for human upper-body animation.

Paperid: 435, https://arxiv.org/pdf/2512.00891.pdf GitHub

Abstract:
Streaming Video Large Language Models (VideoLLMs) have demonstrated impressive performance across various video understanding tasks, but they face significant challenges in real-time deployment due to the high computational cost of processing dense visual tokens from continuous video streams. In streaming video scenarios, the primary bottleneck lies in the Vision Transformer (ViT) encoding stage, where redundant processing of temporally similar frames leads to inefficiency. Additionally, inflated token sequences during LLM pre-filling further exacerbate latency and memory overhead. To address these challenges, we propose \textbf{S}treaming \textbf{T}oken \textbf{C}ompression (\textbf{STC}), a plug-and-play hierarchical framework that seamlessly integrates into existing streaming VideoLLMs, optimizing both ViT encoding and LLM pre-filling stages to accelerate processing. STC introduces two token-level accelerators: \textbf{STC-Cacher}, which reduces ViT encoding overhead by caching and reusing features from temporally similar frames, and \textbf{STC-Pruner}, which compresses the visual token sequence before it enters the LLM, preserving only the most salient tokens based on both spatial and temporal relevance. Extensive experiments on four baseline streaming VideoLLMs across five benchmarks demonstrate that STC outperforms other compression methods. Notably, STC retains up to \textbf{99\%} of accuracy on the ReKV framework while reducing ViT encoding latency and LLM pre-filling latency by \textbf{24.5\%} and \textbf{45.3\%}.

Paperid: 436, https://arxiv.org/pdf/2512.00885.pdf GitHub

Abstract:
Hand-object interaction (HOI) inherently involves dynamics where human manipulations produce distinct spatio-temporal effects on objects. However, existing semantic HOI benchmarks focused either on manipulation or on the resulting effects at a coarse level, lacking fine-grained spatio-temporal reasoning to capture the underlying dynamics in HOI. We introduce HanDyVQA, a fine-grained video question-answering benchmark that comprehensively covers both the manipulation and effect aspects of HOI. HanDyVQA comprises six complementary question types (Action, Process, Objects, Location, State Change, and Object Parts), totalling 11.1K multiple-choice QA pairs. Collected QA pairs recognizing manipulation styles, hand/object motions, and part-level state changes. HanDyVQA also includes 10.3K segmentation masks for Objects and Object Parts questions, enabling the evaluation of object/part-level reasoning in video object segmentation. We evaluated recent video foundation models on our benchmark and found that even the best-performing model, Gemini-2.5-Pro, reached only 73% average accuracy, which is far from human performance (97%). Further analysis shows the remaining challenges in spatial relationship, motion, and part-level geometric understanding. We also found that integrating explicit HOI-related cues into visual features improves performance, offering insights for developing future models with a deeper understanding of HOI dynamics.

Paperid: 437, https://arxiv.org/pdf/2512.00862.pdf GitHub

Abstract:
We introduce HBLLM, a wavelet-enhanced high-fidelity $1$-bit post-training quantization method for Large Language Models (LLMs). By leveraging Haar wavelet transforms to enhance expressive capacity through frequency decomposition, HBLLM significantly improves quantization fidelity while maintaining minimal overhead. This approach features two innovative structure-aware grouping strategies: (1) frequency-aware multi-parameter intra-row grouping and (2) $\ell_2$-norm-based saliency-driven column selection. For non-salient weights, a shared mean is employed across quantization groups within each frequency band to optimize storage efficiency. Experiments conducted on the OPT and LLaMA models demonstrate that HBLLM achieves state-of-the-art performance in $1$-bit quantization, attaining a perplexity of $6.71$ on LLaMA$2$-$13$B with an average weight storage of only $1.08$ bits. Code available at: https://github.com/Yeyke/HBLLM.

Paperid: 438, https://arxiv.org/pdf/2512.00856.pdf GitHub

Abstract:
Probabilistic forecasting is essential for modern risk management, allowing decision-makers to quantify uncertainty in critical systems. This paper tackles this challenge using the volatile REFIT household dataset, which is complicated by a large structural data gap. We first address this by conducting a rigorous comparative experiment to select a Seasonal Imputation method, demonstrating its superiority over linear interpolation in preserving the data's underlying distribution. We then systematically evaluate a hierarchy of models, progressing from classical baselines (SARIMA, Prophet) to machine learning (XGBoost) and advanced deep learning architectures (LSTM). Our findings reveal that classical models fail to capture the data's non-linear, regime-switching behavior. While the LSTM provided the most well-calibrated probabilistic forecast, the Temporal Fusion Transformer (TFT) emerged as the superior all-round model, achieving the best point forecast accuracy (RMSE 481.94) and producing safer, more cautious prediction intervals that effectively capture extreme volatility.

Paperid: 439, https://arxiv.org/pdf/2512.00837.pdf GitHub GitHub

Abstract:
Watermarking acts as a critical safeguard in text generated by Large Language Models (LLMs). By embedding identifiable signals into model outputs, watermarking enables reliable attribution and enhances the security of machine-generated content. Existing approaches typically embed signals by manipulating token generation probabilities. Despite their effectiveness, these methods inherently face a trade-off between detectability and text quality: the signal strength and randomness required for robust watermarking tend to degrade the performance of downstream tasks. In this paper, we design a novel embedding scheme that controls seed pools to facilitate diverse parallel generation of watermarked text. Based on that scheme, we propose WaterSearch, a sentence-level, search-based watermarking framework adaptable to a wide range of existing methods. WaterSearch enhances text quality by jointly optimizing two key aspects: 1) distribution fidelity and 2) watermark signal characteristics. Furthermore, WaterSearch is complemented by a sentence-level detection method with strong attack robustness. We evaluate our method on three popular LLMs across ten diverse tasks. Extensive experiments demonstrate that our method achieves an average performance improvement of 51.01\% over state-of-the-art baselines at a watermark detectability strength of 95\%. In challenging scenarios such as short text generation and low-entropy output generation, our method yields performance gains of 47.78\% and 36.47\%, respectively. Moreover, under different attack senarios including insertion, synonym substitution and paraphrase attasks, WaterSearch maintains high detectability, further validating its robust anti-attack capabilities. Our code is available at \href{https://github.com/Yukang-Lin/WaterSearch}{https://github.com/Yukang-Lin/WaterSearch}.

Paperid: 440, https://arxiv.org/pdf/2512.00832.pdf GitHub GitHub

Abstract:
Panoramic video generation has attracted growing attention due to its applications in virtual reality and immersive media. However, existing methods lack explicit motion control and struggle to generate scenes with large and complex motions. We propose PanFlow, a novel approach that exploits the spherical nature of panoramas to decouple the highly dynamic camera rotation from the input optical flow condition, enabling more precise control over large and dynamic motions. We further introduce a spherical noise warping strategy to promote loop consistency in motion across panorama boundaries. To support effective training, we curate a large-scale, motion-rich panoramic video dataset with frame-level pose and flow annotations. We also showcase the effectiveness of our method in various applications, including motion transfer and video editing. Extensive experiments demonstrate that PanFlow significantly outperforms prior methods in motion fidelity, visual quality, and temporal coherence. Our code, dataset, and models are available at https://github.com/chengzhag/PanFlow.

Paperid: 441, https://arxiv.org/pdf/2512.00831.pdf GitHub

Abstract:
Large Reasoning Models (LRMs) are Large Language Models (LLMs) explicitly trained to generate long-form Chain-of-Thoughts (CoTs), achieving impressive success on challenging tasks like math and programming. However, their underlying reasoning "algorithms" remain poorly understood. To investigate this, we propose ReJump, which represents a reasoning trace as a visitation order over nodes in a tree of intermediate problem-solving steps. Transitions between nodes, which we term jumps, include adjacent moves that capture behaviors such as calculation, and non-adjacent moves that capture behaviors such as backtracking and verification. ReJump enables analyzing LLM reasoning with diverse metrics that quantify exploration, exploitation, overthinking, forgetting, and verification. Using our proposed LLM agent to extract reasoning traces into ReJump format, we evaluate state-of-the-art LRMs on two tasks and find that models with similar accuracy can exhibit distinct reasoning behaviors, while different tasks favor different reasoning styles (e.g., varying balance between exploration and exploitation). To further understand how learning strategies shape reasoning, we use ReJump to compare distilled LRMs with their teachers, CoT-prompted LLMs with LRMs, and to examine how the number of reasoning examples and reinforcement learning affect reasoning behavior. Finally, we show that ReJump can improve reasoning quality at test time through strategies such as ReJump-guided Best-of-N selection and prompt selection. Our code is publicly available at https://github.com/UW-Madison-Lee-Lab/ReJump.

Paperid: 442, https://arxiv.org/pdf/2512.00814.pdf GitHub

Abstract:
Recent advances in post-training paradigms have achieved remarkable success in high-level generation tasks, yet their potential for low-level vision remains rarely explored. Existing image restoration (IR) methods rely on pixel-level hard-fitting to ground-truth images, struggling with over-smoothing and poor generalization. To address these limitations, we propose IRPO, a low-level GRPO-based post-training paradigm that systematically explores both data formulation and reward modeling. We first explore a data formulation principle for low-level post-training paradigm, in which selecting underperforming samples from the pre-training stage yields optimal performance and improved efficiency. Furthermore, we model a reward-level criteria system that balances objective accuracy and human perceptual preference through three complementary components: a General Reward for structural fidelity, an Expert Reward leveraging Qwen-VL for perceptual alignment, and a Restoration Reward for task-specific low-level quality. Comprehensive experiments on six in-domain and five out-of-domain (OOD) low-level benchmarks demonstrate that IRPO achieves state-of-the-art results across diverse degradation types, surpassing the AdaIR baseline by 0.83 dB on in-domain tasks and 3.43 dB on OOD settings. Our code can be shown in https://github.com/HaoxuanXU1024/IRPO.

Paperid: 443, https://arxiv.org/pdf/2512.00789.pdf GitHub

Abstract:
Token sampling strategies critically influence text generation quality in large language models (LLMs). However, existing methods introduce additional hyperparameters, requiring extensive tuning and complicating deployment. We present Entropy Equilibrium Sampling (EES), an auxiliary hyperparameter-free approach inspired by information theory that can dynamically adjust candidate sets by balancing normalized entropy with probability mass. We evaluate EES on both reasoning and generation tasks across a range of model architectures. Our results show that EES consistently performs well across temperature settings, delivering competitive accuracy and coherence while maintaining diversity. By eliminating the need for hyperparameter tuning, EES greatly simplifies deployment while improving performance. Code is available at https://github.com/shuanncai/EES

Paperid: 444, https://arxiv.org/pdf/2512.00762.pdf GitHub GitHub

Abstract:
A longstanding goal in computer vision is to model motions from videos, while the representations behind motions, i.e. the invisible physical interactions that cause objects to deform and move, remain largely unexplored. In this paper, we study how to recover the invisible forces from visual observations, e.g., estimating the wind field by observing a leaf falling to the ground. Our key innovation is an end-to-end differentiable inverse graphics framework, which jointly models object geometry, physical properties, and interactions directly from videos. Through backpropagation, our approach enables the recovery of force representations from object motions. We validate our method on both synthetic and real-world scenarios, and the results demonstrate its ability to infer plausible force fields from videos. Furthermore, we show the potential applications of our approach, including physics-based video generation and editing. We hope our approach sheds light on understanding and modeling the physical process behind pixels, bridging the gap between vision and physics. Please check more video results in our \href{https://chaoren2357.github.io/seeingthewind/}{project page}.

Paperid: 445, https://arxiv.org/pdf/2512.00752.pdf GitHub

Abstract:
Generative models, such as diffusion and autoregressive approaches, have demonstrated impressive capabilities in editing natural images. However, applying these tools to scientific charts rests on a flawed assumption: a chart is not merely an arrangement of pixels but a visual representation of structured data governed by a graphical grammar. Consequently, chart editing is not a pixel-manipulation task but a structured transformation problem. To address this fundamental mismatch, we introduce \textit{FigEdit}, a large-scale benchmark for scientific figure editing comprising over 30,000 samples. Grounded in real-world data, our benchmark is distinguished by its diversity, covering 10 distinct chart types and a rich vocabulary of complex editing instructions. The benchmark is organized into five distinct and progressively challenging tasks: single edits, multi edits, conversational edits, visual-guidance-based edits, and style transfer. Our evaluation of a range of state-of-the-art models on this benchmark reveals their poor performance on scientific figures, as they consistently fail to handle the underlying structured transformations required for valid edits. Furthermore, our analysis indicates that traditional evaluation metrics (e.g., SSIM, PSNR) have limitations in capturing the semantic correctness of chart edits. Our benchmark demonstrates the profound limitations of pixel-level manipulation and provides a robust foundation for developing and evaluating future structure-aware models. By releasing \textit{FigEdit} (https://github.com/adobe-research/figure-editing), we aim to enable systematic progress in structure-aware figure editing, provide a common ground for fair comparison, and encourage future research on models that understand both the visual and semantic layers of scientific charts.

Paperid: 446, https://arxiv.org/pdf/2512.00748.pdf GitHub

Abstract:
Medical image segmentation is inherently influenced by data uncertainty, arising from ambiguous boundaries in medical scans and inter-observer variability in diagnosis. To address this challenge, previous works formulated the multi-rater medical image segmentation task, where multiple experts provide separate annotations for each image. However, existing models are typically constrained to either generate diverse segmentation that lacks expert specificity or to produce personalized outputs that merely replicate individual annotators. We propose Probabilistic modeling of multi-rater medical image Segmentation (ProSeg) that simultaneously enables both diversification and personalization. Specifically, we introduce two latent variables to model expert annotation preferences and image boundary ambiguity. Their conditional probabilistic distributions are then obtained through variational inference, allowing segmentation outputs to be generated by sampling from these distributions. Extensive experiments on both the nasopharyngeal carcinoma dataset (NPC) and the lung nodule dataset (LIDC-IDRI) demonstrate that our ProSeg achieves a new state-of-the-art performance, providing segmentation results that are both diverse and expert-personalized. Code can be found in https://github.com/AI4MOL/ProSeg.

Paperid: 447, https://arxiv.org/pdf/2512.00743.pdf GitHub GitHub

Abstract:
Recently, Group Relative Policy Optimization (GRPO) has shown promising potential for aligning text-to-image (T2I) models, yet existing GRPO-based methods suffer from two critical limitations. (1) \textit{Shared credit assignment}: trajectory-level advantages derived from group-normalized sparse terminal rewards are uniformly applied across timesteps, failing to accurately estimate the potential of early denoising steps with vast exploration spaces. (2) \textit{Reward-mixing}: predefined weights for combining multi-objective rewards (e.g., text accuracy, visual quality, text color)--which have mismatched scales and variances--lead to unstable gradients and conflicting updates. To address these issues, we propose \textbf{Multi-GRPO}, a multi-group advantage estimation framework with two orthogonal grouping mechanisms. For better credit assignment, we introduce tree-based trajectories inspired by Monte Carlo Tree Search: branching trajectories at selected early denoising steps naturally forms \emph{temporal groups}, enabling accurate advantage estimation for early steps via descendant leaves while amortizing computation through shared prefixes. For multi-objective optimization, we introduce \emph{reward-based grouping} to compute advantages for each reward function \textit{independently} before aggregation, disentangling conflicting signals. To facilitate evaluation of multiple objective alignment, we curate \textit{OCR-Color-10}, a visual text rendering dataset with explicit color constraints. Across the single-reward \textit{PickScore-25k} and multi-objective \textit{OCR-Color-10} benchmarks, Multi-GRPO achieves superior stability and alignment performance, effectively balancing conflicting objectives. Code will be publicly available at \href{https://github.com/fikry102/Multi-GRPO}{https://github.com/fikry102/Multi-GRPO}.

Paperid: 448, https://arxiv.org/pdf/2512.00727.pdf GitHub

Abstract:
Reinforcement learning has recently enabled impressive locomotion capabilities on legged robots; however, most policy architectures remain morphology- and symmetry-agnostic, leading to inefficient training and limited generalization. This work introduces MS-PPO, a morphological-symmetry-equivariant policy learning framework that encodes robot kinematic structure and morphological symmetries directly into the policy network. We construct a morphology-informed graph neural architecture that is provably equivariant with respect to the robot's morphological symmetry group actions, ensuring consistent policy responses under symmetric states while maintaining invariance in value estimation. This design eliminates the need for tedious reward shaping or costly data augmentation, which are typically required to enforce symmetry. We evaluate MS-PPO in simulation on Unitree Go2 and Xiaomi CyberDog2 robots across diverse locomotion tasks, including trotting, pronking, slope walking, and bipedal turning, and further deploy the learned policies on hardware. Extensive experiments show that MS-PPO achieves superior training stability, symmetry generalization ability, and sample efficiency in challenging locomotion tasks, compared to state-of-the-art baselines. These findings demonstrate that embedding both kinematic structure and morphological symmetry into policy learning provides a powerful inductive bias for legged robot locomotion control. Our code will be made publicly available at https://lunarlab-gatech.github.io/MS-PPO/.

Paperid: 449, https://arxiv.org/pdf/2512.00716.pdf GitHub

Abstract:
Covariate distribution shift occurs when certain structural features present in the test set are absent from the training set. It is a common type of out-of-distribution (OOD) problem, frequently encountered in real-world graph data with complex structures. Existing research has revealed that most out-of-the-box graph neural networks (GNNs) fail to account for covariate shifts. Furthermore, we observe that existing methods aimed at addressing covariate shifts often fail to fully leverage the rich information contained within the latent space. Motivated by the potential of the latent space, we introduce a new method called MPAIACL for More Powerful Adversarial Invariant Augmentation using Contrastive Learning. MPAIACL leverages contrastive learning to unlock the full potential of vector representations by harnessing their intrinsic information. Through extensive experiments, MPAIACL demonstrates its robust generalization and effectiveness, as it performs well compared with other baselines across various public OOD datasets. The code is publicly available at https://github.com/flzeng1/MPAIACL.

Paperid: 450, https://arxiv.org/pdf/2512.00708.pdf GitHub

Abstract:
Advancements in AI for science unlocks capabilities for critical drug discovery tasks such as protein-ligand binding affinity prediction. However, current models overfit to existing oversimplified datasets that does not represent naturally occurring and biologically relevant proteins with modifications. In this work, we curate a complete and modification-aware version of the widely used DAVIS dataset by incorporating 4,032 kinase-ligand pairs involving substitutions, insertions, deletions, and phosphorylation events. This enriched dataset enables benchmarking of predictive models under biologically realistic conditions. Based on this new dataset, we propose three benchmark settings-Augmented Dataset Prediction, Wild-Type to Modification Generalization, and Few-Shot Modification Generalization-designed to assess model robustness in the presence of protein modifications. Through extensive evaluation of both docking-free and docking-based methods, we find that docking-based model generalize better in zero-shot settings. In contrast, docking-free models tend to overfit to wild-type proteins and struggle with unseen modifications but show notable improvement when fine-tuned on a small set of modified examples. We anticipate that the curated dataset and benchmarks offer a valuable foundation for developing models that better generalize to protein modifications, ultimately advancing precision medicine in drug discovery. The benchmark is available at: https://github.com/ZhiGroup/DAVIS-complete

Paperid: 451, https://arxiv.org/pdf/2512.00705.pdf GitHub

Abstract:
Dynamic random walks are fundamental to various graph analysis applications, offering advantages by adapting to evolving graph properties. Their runtime-dependent transition probabilities break down the pre-computation strategy that underpins most existing CPU and GPU static random walk optimizations. This leaves practitioners suffering from suboptimal frameworks and having to write hand-tuned kernels that do not adapt to workload diversity. To handle this issue, we present FlexiWalker, the first GPU framework that delivers efficient, workload-generic support for dynamic random walks. Our design-space study shows that rejection sampling and reservoir sampling are more suitable than other sampling techniques under massive parallelism. Thus, we devise (i) new high-performance kernels for them that eliminate global reductions, redundant memory accesses, and random-number generation. Given the necessity of choosing the best-fitting sampling strategy at runtime, we adopt (ii) a lightweight first-order cost model that selects the faster kernel per node at runtime. To enhance usability, we introduce (iii) a compile-time component that automatically specializes user-supplied walk logic into optimized building blocks. On various dynamic random walk workloads with real-world graphs, FlexiWalker outperforms the best published CPU/GPU baselines by geometric means of 73.44x and 5.91x, respectively, while successfully executing workloads that prior systems cannot support. We open-source FlexiWalker in https://github.com/AIS-SNU/FlexiWalker.

Paperid: 452, https://arxiv.org/pdf/2512.00700.pdf GitHub

Abstract:
We propose a new neural network architecture called CAR-net (CAscade Refinement Network) to deblur images that are subject to rotational motion blur. Our architecture is specifically designed for the semi-blind scenarios where only noisy information of the rotational motion blur angle is available. The core of our approach is progressive refinement process that starts with an initial deblurred estimate obtained from frequency-domain inversion; A series of refinement stages take the current deblurred image to predict and apply residual correction to the current estimate, progressively suppressing artifacts and restoring fine details. To handle parameter uncertainty, our architecture accommodates an optional angle detection module which can be trained end-to-end with refinement modules. We provide a detailed description of our architecture and illustrate its efficiency through experiments using both synthetic and real-life images. Our code and model as well as the links to the datasets are available at https://github.com/tony123105/CAR-Net

Paperid: 453, https://arxiv.org/pdf/2512.00691.pdf GitHub

Abstract:
Gait patterns play a critical role in human identification and healthcare analytics, yet current progress remains constrained by small, narrowly designed models that fail to scale or generalize. Building a unified gait foundation model requires addressing two longstanding barriers: (a) Scalability. Why have gait models historically failed to follow scaling laws? (b) Generalization. Can one model serve the diverse gait tasks that have traditionally been studied in isolation? We introduce FoundationGait, the first scalable, self-supervised pretraining framework for gait understanding. Its largest version has nearly 0.13 billion parameters and is pretrained on 12 public gait datasets comprising over 2 million walking sequences. Extensive experiments demonstrate that FoundationGait, with or without fine-tuning, performs robustly across a wide spectrum of gait datasets, conditions, tasks (e.g., human identification, scoliosis screening, depression prediction, and attribute estimation), and even input modality. Notably, it achieves 48.0% zero-shot rank-1 accuracy on the challenging in-the-wild Gait3D dataset (1,000 test subjects) and 64.5% on the largest in-the-lab OU-MVLP dataset (5,000+ test subjects), setting a new milestone in robust gait recognition. Coming code and model: https://github.com/ShiqiYu/OpenGait.

Paperid: 454, https://arxiv.org/pdf/2512.00677.pdf GitHub

Abstract:
Recent progress in 4D representations, such as Dynamic NeRF and 4D Gaussian Splatting (4DGS), has enabled dynamic 4D scene reconstruction. However, text-driven 4D scene editing remains under-explored due to the challenge of ensuring both multi-view and temporal consistency across space and time during editing. Existing studies rely on 2D diffusion models that edit frames independently, often causing motion distortion, geometric drift, and incomplete editing. We introduce Dynamic-eDiTor, a training-free text-driven 4D editing framework leveraging Multimodal Diffusion Transformer (MM-DiT) and 4DGS. This mechanism consists of Spatio-Temporal Sub-Grid Attention (STGA) for locally consistent cross-view and temporal fusion, and Context Token Propagation (CTP) for global propagation via token inheritance and optical-flow-guided token replacement. Together, these components allow Dynamic-eDiTor to perform seamless, globally consistent multi-view video without additional training and directly optimize pre-trained source 4DGS. Extensive experiments on multi-view video dataset DyNeRF demonstrate that our method achieves superior editing fidelity and both multi-view and temporal consistency prior approaches. Project page for results and code: https://di-lee.github.io/dynamic-eDiTor/

Paperid: 455, https://arxiv.org/pdf/2512.00564.pdf GitHub

Abstract:
A key aspect of learned partial differential equation (PDE) solvers is that the main cost often comes from generating training data with classical solvers rather than learning the model itself. Another is that there are clear axes of difficulty--e.g., more complex geometries and higher Reynolds numbers--along which problems become (1) harder for classical solvers and thus (2) more likely to benefit from neural speedups. Towards addressing this chicken-and-egg challenge, we study difficulty transfer on 2D incompressible Navier-Stokes, systematically varying task complexity along geometry (number and placement of obstacles), physics (Reynolds number), and their combination. Similar to how it is possible to spend compute to pre-train foundation models and improve their performance on downstream tasks, we find that by classically solving (analogously pre-generating) many low and medium difficulty examples and including them in the training set, it is possible to learn high-difficulty physics from far fewer samples. Furthermore, we show that by combining low and high difficulty data, we can spend 8.9x less compute on pre-generating a dataset to achieve the same error as using only high difficulty examples. Our results highlight that how we allocate classical-solver compute across difficulty levels is as important as how much we allocate overall, and suggest substantial gains from principled curation of pre-generated PDE data for neural solvers. Our code is available at https://github.com/Naman-Choudhary-AI-ML/pregenerating-pde

Paperid: 456, https://arxiv.org/pdf/2512.00532.pdf GitHub GitHub

Abstract:
Generating realistic robotic manipulation videos is an important step toward unifying perception, planning, and action in embodied agents. While existing video diffusion models require large domain-specific datasets and struggle to generalize, recent image generation models trained on language-image corpora exhibit strong compositionality, including the ability to synthesize temporally coherent grid images. This suggests a latent capacity for video-like generation even without explicit temporal modeling. We explore whether such models can serve as visual planners for robots when lightly adapted using LoRA finetuning. We propose a two-part framework that includes: (1) text-conditioned generation, which uses a language instruction and the first frame, and (2) trajectory-conditioned generation, which uses a 2D trajectory overlay and the same initial frame. Experiments on the Jaco Play dataset, Bridge V2, and the RT1 dataset show that both modes produce smooth, coherent robot videos aligned with their respective conditions. Our findings indicate that pretrained image generators encode transferable temporal priors and can function as video-like robotic planners under minimal supervision. Code is released at \href{https://github.com/pangye202264690373/Image-Generation-as-a-Visual-Planner-for-Robotic-Manipulation}{https://github.com/pangye202264690373/Image-Generation-as-a-Visual-Planner-for-Robotic-Manipulation}.

Paperid: 457, https://arxiv.org/pdf/2512.00504.pdf GitHub

Abstract:
Recent reasoning large language models (LLMs) excel in complex tasks but encounter significant computational and memory challenges due to long sequence lengths. KV cache compression has emerged as an effective approach to greatly enhance the efficiency of reasoning. However, existing methods often focus on prompt compression or token eviction with local attention score, overlooking the long-term importance of tokens. We propose G-KV, a KV cache eviction method that employs a global scoring mechanism, combining local and historical attention scores to more accurately assess token importance. Additionally, we introduce post-training techniques, including reinforcement learning and distillation, to optimize models for compressed KV cache settings. The code of this paper is available on: https://github.com/microsoft/G-KV.

Paperid: 458, https://arxiv.org/pdf/2512.00473.pdf GitHub

Abstract:
With the continuous advancement of image generation technology, advanced models such as GPT-Image-1 and Qwen-Image have achieved remarkable text-to-image consistency and world knowledge However, these models still fall short in photorealistic image generation. Even on simple T2I tasks, they tend to produce " fake" images with distinct AI artifacts, often characterized by "overly smooth skin" and "oily facial sheens". To recapture the original goal of "indistinguishable-from-reality" generation, we propose RealGen, a photorealistic text-to-image framework. RealGen integrates an LLM component for prompt optimization and a diffusion model for realistic image generation. Inspired by adversarial generation, RealGen introduces a "Detector Reward" mechanism, which quantifies artifacts and assesses realism using both semantic-level and feature-level synthetic image detectors. We leverage this reward signal with the GRPO algorithm to optimize the entire generation pipeline, significantly enhancing image realism and detail. Furthermore, we propose RealBench, an automated evaluation benchmark employing Detector-Scoring and Arena-Scoring. It enables human-free photorealism assessment, yielding results that are more accurate and aligned with real user experience. Experiments demonstrate that RealGen significantly outperforms general models like GPT-Image-1 and Qwen-Image, as well as specialized photorealistic models like FLUX-Krea, in terms of realism, detail, and aesthetics. The code is available at https://github.com/yejy53/RealGen.

Paperid: 459, https://arxiv.org/pdf/2512.00470.pdf GitHub

Abstract:
Diffusion models have demonstrated strong capabilities for modeling human-like driving behaviors in autonomous driving, but their iterative sampling process induces substantial latency, and operating directly on raw trajectory points forces the model to spend capacity on low-level kinematics, rather than high-level multi-modal semantics. To address these limitations, we propose LAtent Planner (LAP), a framework that plans in a VAE-learned latent space that disentangles high-level intents from low-level kinematics, enabling our planner to capture rich, multi-modal driving strategies. We further introduce a fine-grained feature distillation mechanism to guide a better interaction and fusion between the high-level semantic planning space and the vectorized scene context. Notably, LAP can produce high-quality plans in one single denoising step, substantially reducing computational overhead. Through extensive evaluations on the large-scale nuPlan benchmark, LAP achieves state-of-the-art closed-loop performance among learning-based planning methods, while demonstrating an inference speed-up of at most 10 times over previous SOTA approaches. Code will be released at: https://github.com/jhz1192/Latent-Planner.

Paperid: 460, https://arxiv.org/pdf/2512.00451.pdf GitHub

Abstract:
Voice communication in bandwidth-constrained environments--maritime, satellite, and tactical networks--remains prohibitively expensive. Traditional codecs struggle below 1 kbps, while existing semantic approaches (STT-TTS) sacrifice prosody and speaker identity. We present STCTS, a generative semantic compression framework enabling natural voice communication at 80 bps. STCTS explicitly decomposes speech into linguistic content, prosodic expression, and speaker timbre, applying tailored compression: context-aware text encoding (70 bps), sparse prosody transmission via TTS interpolation (<14 bps at 0.1-1 Hz), and amortized speaker embedding. Evaluations on LibriSpeech demonstrate a 75x bitrate reduction versus Opus (6 kbps) and 12x versus EnCodec (1 kbps), while maintaining perceptual quality (NISQA MOS > 4.26), graceful degradation under packet loss and noise resilience. We also discover a bimodal quality distribution with prosody sampling rate: sparse and dense updates both achieve high quality, while mid-range rates degrade due to perceptual discontinuities--guiding optimal configuration design. Beyond efficiency, our modular architecture supports privacy-preserving encryption, human-interpretable transmission, and flexible deployment on edge devices, offering a robust solution for ultra-low bandwidth scenarios.

Paperid: 461, https://arxiv.org/pdf/2512.00438.pdf GitHub GitHub

Abstract:
Test-time scaling (TTS) has become a prevalent technique in image generation, significantly boosting output quality by expanding the number of parallel samples and filtering them using pre-trained reward models. However, applying this powerful methodology to the next-token prediction (NTP) paradigm remains challenging. The primary obstacle is the low correlation between the reward of an image decoded from an intermediate token sequence and the reward of the fully generated image. Consequently, these incomplete intermediate representations prove to be poor indicators for guiding the pruning direction, a limitation that stems from their inherent incompleteness in scale or semantic content. To effectively address this critical issue, we introduce the Filling-Based Reward (FR). This novel design estimates the approximate future trajectory of an intermediate sample by finding and applying a reasonable filling scheme to complete the sequence. Both the correlation coefficient between rewards of intermediate samples and final samples, as well as multiple intrinsic signals like token confidence, indicate that the FR provides an excellent and reliable metric for accurately evaluating the quality of intermediate samples. Building upon this foundation, we propose FR-TTS, a sophisticated scaling strategy. FR-TTS efficiently searches for good filling schemes and incorporates a diversity reward with a dynamic weighting schedule to achieve a balanced and comprehensive evaluation of intermediate samples. We experimentally validate the superiority of FR-TTS over multiple established benchmarks and various reward models. Code is available at \href{https://github.com/xuhang07/FR-TTS}{https://github.com/xuhang07/FR-TTS}.

Paperid: 462, https://arxiv.org/pdf/2512.00425.pdf GitHub

Abstract:
Recent video diffusion models can synthesize visually compelling clips, yet often violate basic physical laws-objects float, accelerations drift, and collisions behave inconsistently-revealing a persistent gap between visual realism and physical realism. We propose $\texttt{NewtonRewards}$, the first physics-grounded post-training framework for video generation based on $\textit{verifiable rewards}$. Instead of relying on human or VLM feedback, $\texttt{NewtonRewards}$ extracts $\textit{measurable proxies}$ from generated videos using frozen utility models: optical flow serves as a proxy for velocity, while high-level appearance features serve as a proxy for mass. These proxies enable explicit enforcement of Newtonian structure through two complementary rewards: a Newtonian kinematic constraint enforcing constant-acceleration dynamics, and a mass conservation reward preventing trivial, degenerate solutions. We evaluate $\texttt{NewtonRewards}$ on five Newtonian Motion Primitives (free fall, horizontal/parabolic throw, and ramp sliding down/up) using our newly constructed large-scale benchmark, $\texttt{NewtonBench-60K}$. Across all primitives in visual and physics metrics, $\texttt{NewtonRewards}$ consistently improves physical plausibility, motion smoothness, and temporal coherence over prior post-training methods. It further maintains strong performance under out-of-distribution shifts in height, speed, and friction. Our results show that physics-grounded verifiable rewards offer a scalable path toward physics-aware video generation.

Paperid: 463, https://arxiv.org/pdf/2512.00387.pdf GitHub

Abstract:
Recent image editing models boast next-level intelligent capabilities, facilitating cognition- and creativity-informed image editing. Yet, existing benchmarks provide too narrow a scope for evaluation, failing to holistically assess these advanced abilities. To address this, we introduce WiseEdit, a knowledge-intensive benchmark for comprehensive evaluation of cognition- and creativity-informed image editing, featuring deep task depth and broad knowledge breadth. Drawing an analogy to human cognitive creation, WiseEdit decomposes image editing into three cascaded steps, i.e., Awareness, Interpretation, and Imagination, each corresponding to a task that poses a challenge for models to complete at the specific step. It also encompasses complex tasks, where none of the three steps can be finished easily. Furthermore, WiseEdit incorporates three fundamental types of knowledge: Declarative, Procedural, and Metacognitive knowledge. Ultimately, WiseEdit comprises 1,220 test cases, objectively revealing the limitations of SoTA image editing models in knowledge-based cognitive reasoning and creative composition capabilities. The benchmark, evaluation code, and the generated images of each model will be made publicly available soon. Project Page: https://qnancy.github.io/wiseedit_project_page/.

Paperid: 464, https://arxiv.org/pdf/2512.00343.pdf GitHub

Abstract:
Vision-language pretrained models (VLPs) such as CLIP have achieved remarkable success, but are also highly vulnerable to backdoor attacks. Given a model fine-tuned by an untrusted third party, determining whether the model has been injected with a backdoor is a critical and challenging problem. Existing detection methods usually rely on prior knowledge of training dataset, backdoor triggers and targets, or downstream classifiers, which may be impractical for real-world applications. To address this, To address this challenge, we introduce Assimilation Matters in DETection (AMDET), a novel model-level detection framework that operates without any such prior knowledge. Specifically, we first reveal the feature assimilation property in backdoored text encoders: the representations of all tokens within a backdoor sample exhibit a high similarity. Further analysis attributes this effect to the concentration of attention weights on the trigger token. Leveraging this insight, AMDET scans a model by performing gradient-based inversion on token embeddings to recover implicit features that capable of activating backdoor behaviors. Furthermore, we identify the natural backdoor feature in the OpenAI's official CLIP model, which are not intentionally injected but still exhibit backdoor-like behaviors. We then filter them out from real injected backdoor by analyzing their loss landscapes. Extensive experiments on 3,600 backdoored and benign-finetuned models with two attack paradigms and three VLP model structures show that AMDET detects backdoors with an F1 score of 89.90%. Besides, it achieves one complete detection in approximately 5 minutes on a RTX 4090 GPU and exhibits strong robustness against adaptive attacks. Code is available at: https://github.com/Robin-WZQ/AMDET

Paperid: 465, https://arxiv.org/pdf/2512.00336.pdf GitHub

Abstract:
The rapid advancement of AI-generated multimodal video-audio content has raised significant concerns regarding information security and content authenticity. Existing synthetic video datasets predominantly focus on the visual modality alone, while the few incorporating audio are largely confined to facial deepfakes--a limitation that fails to address the expanding landscape of general multimodal AI-generated content and substantially impedes the development of trustworthy detection systems. To bridge this critical gap, we introduce the Multimodal Video-Audio Dataset (MVAD), the first comprehensive dataset specifically designed for detecting AI-generated multimodal video-audio content. Our dataset exhibits three key characteristics: (1) genuine multimodality with samples generated according to three realistic video-audio forgery patterns; (2) high perceptual quality achieved through diverse state-of-the-art generative models; and (3) comprehensive diversity spanning realistic and anime visual styles, four content categories (humans, animals, objects, and scenes), and four video-audio multimodal data types. Our dataset will be available at https://github.com/HuMengXue0104/MVAD.

Paperid: 466, https://arxiv.org/pdf/2512.00333.pdf GitHub

Abstract:
While large language models excel on high-resource multilingual tasks, low- and extremely low-resource Indic languages remain severely under-evaluated. We present IndicParam, a human-curated benchmark of over 13,000 multiple-choice questions covering 11 such languages (Nepali, Gujarati, Marathi, Odia as low-resource; Dogri, Maithili, Rajasthani, Sanskrit, Bodo, Santali, Konkani as extremely low-resource) plus Sanskrit-English code-mixed set. We evaluated 19 LLMs, both proprietary and open-weights, which reveals that even the top-performing GPT-5 reaches only 45.0% average accuracy, followed by DeepSeek-3.2 (43.1) and Claude-4.5 (42.7). We additionally label each question as knowledge-oriented or purely linguistic to discriminate factual recall from grammatical proficiency. Further, we assess the ability of LLMs to handle diverse question formats-such as list-based matching, assertion-reason pairs, and sequence ordering-alongside conventional multiple-choice questions. IndicParam provides insights into limitations of cross-lingual transfer and establishes a challenging benchmark for Indic languages. The dataset is available at https://huggingface.co/datasets/bharatgenai/IndicParam. Scripts to run benchmark are present at https://github.com/ayushbits/IndicParam.

Paperid: 467, https://arxiv.org/pdf/2512.00310.pdf GitHub

Abstract:
Unsupervised anomaly detection aims to identify anomalies without pixel-level annotations. Synthetic anomaly-based methods exhibit a unique capacity to introduce controllable irregularities with known masks, enabling explicit supervision during training. However, existing methods often produce synthetic anomalies that are visually distinct from real pathological patterns and ignore anatomical structure. This paper presents a novel Anatomy-aware Realistic Texture-based Anomaly Synthesis framework (ART-ASyn) for chest X-rays that generates realistic and anatomically consistent lung opacity related anomalies using texture-based augmentation guided by our proposed Progressive Binary Thresholding Segmentation method (PBTSeg) for lung segmentation. The generated paired samples of synthetic anomalies and their corresponding precise pixel-level anomaly mask for each normal sample enable explicit segmentation supervision. In contrast to prior work limited to one-class classification, ART-ASyn is further evaluated for zero-shot anomaly segmentation, demonstrating generalizability on an unseen dataset without target-domain annotations. Code availability is available at https://github.com/angelacao-hub/ART-ASyn.

Paperid: 468, https://arxiv.org/pdf/2512.00288.pdf GitHub

Abstract:
Benchmarking is central to optimization research, yet existing test suites for continuous optimization remain limited: classical collections are fixed and rigid, while previous generators cover only narrow families of landscapes with restricted variability and control over details. This paper introduces PORTAL (Platform for Optimization Research, Testing, Analysis, and Learning), a general benchmark generator that provides fine-grained, independent control over basin curvature, conditioning, variable interactions, and surface ruggedness. PORTAL's layered design spans from individual components to block-wise compositions of multi-component landscapes with controllable partial separability and imbalanced block contributions. It offers precise control over the shape of each component in every dimension and direction, and supports diverse transformation patterns through both element-wise and coupling operators with compositional sequencing. All transformations preserve component centers and local quadratic structure, ensuring stability and interpretability. A principled neutralization mechanism prevents unintended component domination caused by exponent or scale disparities, which addresses a key limitation of prior landscape generators. On this foundation, transformations introduce complex landscape characteristics, such as multimodality, asymmetry, and heterogeneous ruggedness, in a controlled and systematic way. PORTAL enables systematic algorithm analysis by supporting both isolation of specific challenges and progressive difficulty scaling. It also facilitates the creation of diverse datasets for meta-algorithmic research, tailored benchmark suite design, and interactive educational use. The complete Python and MATLAB source code for PORTAL is publicly available at [https://github.com/EvoMindLab/PORTAL].

Paperid: 469, https://arxiv.org/pdf/2512.00269.pdf GitHub

Abstract:
Understanding the relationship between pathological and healthy brain structures is fundamental to neuroimaging, connecting disease diagnosis and detection with modeling, prediction, and treatment planning. However, paired pathological-healthy data are extremely difficult to obtain, as they rely on pre- and post-treatment imaging, constrained by clinical outcomes and longitudinal data availability. Consequently, most existing brain image generation and editing methods focus on visual quality yet remain domain-specific, treating pathological and healthy image modeling independently. We introduce USB (Unified Synthetic Brain), the first end-to-end framework that unifies bidirectional generation and editing of pathological and healthy brain images. USB models the joint distribution of lesions and brain anatomy through a paired diffusion mechanism and achieves both pathological and healthy image generation. A consistency guidance algorithm further preserves anatomical consistency and lesion correspondence during bidirectional pathology-healthy editing. Extensive experiments on six public brain MRI datasets including healthy controls, stroke, and Alzheimer's patients, demonstrate USB's ability to produce diverse and realistic results. By establishing the first unified benchmark for brain image generation and editing, USB opens opportunities for scalable dataset creation and robust neuroimaging analysis. Code is available at https://github.com/jhuldr/USB.

Paperid: 470, https://arxiv.org/pdf/2512.00264.pdf GitHub

Abstract:
We present the first geometric deep learning framework based on point cloud representation for 3D four-chamber cardiac reconstruction from cine MRI data. This work addresses a long-standing limitation in conventional cine MRI, which typically provides only 2D slice images of the heart, thereby restricting a comprehensive understanding of cardiac morphology and physiological mechanisms in both healthy and pathological conditions. To overcome this, we propose \textbf{HeartFormer}, a novel point cloud completion network that extends traditional single-class point cloud completion to the multi-class. HeartFormer consists of two key components: a Semantic-Aware Dual-Structure Transformer Network (SA-DSTNet) and a Semantic-Aware Geometry Feature Refinement Transformer Network (SA-GFRTNet). SA-DSTNet generates an initial coarse point cloud with both global geometry features and substructure geometry features. Guided by these semantic-geometry representations, SA-GFRTNet progressively refines the coarse output, effectively leveraging both global and substructure geometric priors to produce high-fidelity and geometrically consistent reconstructions. We further construct \textbf{HeartCompv1}, the first publicly available large-scale dataset with 17,000 high-resolution 3D multi-class cardiac meshes and point-clouds, to establish a general benchmark for this emerging research direction. Extensive cross-domain experiments on HeartCompv1 and UK Biobank demonstrate that HeartFormer achieves robust, accurate, and generalizable performance, consistently surpassing state-of-the-art (SOTA) methods. Code and dataset will be released upon acceptance at: https://github.com/10Darren/HeartFormer.

Paperid: 471, https://arxiv.org/pdf/2512.00234.pdf GitHub

Abstract:
There has been significant progress in open-source text-only translation large language models (LLMs) with better language coverage and quality. However, these models can be only used in cascaded pipelines for speech translation (ST), performing automatic speech recognition first followed by translation. This introduces additional latency, which is particularly critical in simultaneous ST (SimulST), and prevents the model from exploiting multimodal context, such as images, which can aid disambiguation. Pretrained multimodal foundation models (MMFMs) already possess strong perception and reasoning capabilities across multiple modalities, but generally lack the multilingual coverage and specialized translation performance of dedicated translation LLMs. To build an effective multimodal translation system, we propose an end-to-end approach that fuses MMFMs with translation LLMs. We introduce a novel fusion strategy that connects hidden states from multiple layers of a pretrained MMFM to a translation LLM, enabling joint end-to-end training. The resulting model, OmniFusion, built on Omni 2.5-7B as the MMFM and SeedX PPO-7B as the translation LLM, can perform speech-to-text, speech-and-image-to-text, and text-and-image-to-text translation. Experiments demonstrate that OmniFusion effectively leverages both audio and visual inputs, achieves a 1-second latency reduction in SimulST compared to cascaded pipelines and also improves the overall translation quality\footnote{Code is available at https://github.com/saikoneru/OmniFusion}.

Paperid: 472, https://arxiv.org/pdf/2512.00218.pdf GitHub

Abstract:
AI systems that output their reasoning in natural language offer an opportunity for safety -- we can \emph{monitor} their chain of thought (CoT) for undesirable reasoning, such as the pursuit of harmful objectives. However, the extent to which CoT faithfully reflects the underlying reasoning process, and hence the extent to which it can be usefully monitored, may be influenced by certain aspects of training. We investigate how different \emph{training incentives}, applied to a reasoning model, affect its monitorability. We introduce a novel methodology for measuring monitorability according to whether a monitor can predict a key latent variable using the model's reasoning. When controlling for accuracy, we do not find evidence for consistent effects from commonly used incentives (length penalties and KL regularisation), but we find that adversarial optimisation (penalising monitor accuracy) degrades monitor performance, while direct optimisation for monitorability does not reliably lead to improvements. Our code is available at https://github.com/QiyaoWei/reasoning-under-pressure.

Paperid: 473, https://arxiv.org/pdf/2512.00181.pdf GitHub

Abstract:
Tabular data drive most real-world machine learning applications, yet building general-purpose models for them remains difficult. Mixed numeric and categorical fields, weak feature structure, and limited labeled data make scaling and generalization challenging. To this end, we introduce Orion-Bix, a tabular foundation model that combines biaxial attention with meta-learned in-context reasoning for few-shot tabular learning. Its encoder alternates standard, grouped, hierarchical, and relational attention, fusing their outputs through multi-CLS summarization to capture both local and global dependencies efficiently. A label-aware ICL head adapts on the fly and scales to large label spaces via hierarchical decision routing. Meta-trained on synthetically generated, structurally diverse tables with causal priors, Orion-Bix learns transferable inductive biases across heterogeneous data. Delivered as a scikit-learn compatible foundation model, it outperforms gradient-boosting baselines and remains competitive with state-of-the-art tabular foundation models on public benchmarks, showing that biaxial attention with episodic meta-training enables robust, few-shot-ready tabular learning. The model is publicly available at https://github.com/Lexsi-Labs/Orion-BiX .

Paperid: 474, https://arxiv.org/pdf/2512.00130.pdf GitHub

Abstract:
Cutmix-based data augmentation, which uses a cut-and-paste strategy, has shown remarkable generalization capabilities in deep learning. However, existing methods primarily consider global semantics with image-level constraints, which excessively reduces attention to the discriminative local context of the class and leads to a performance improvement bottleneck. Moreover, existing methods for generating augmented samples usually involve cutting and pasting rectangular or square regions, resulting in a loss of object part information. To mitigate the problem of inconsistency between the augmented image and the generated mixed label, existing methods usually require double forward propagation or rely on an external pre-trained network for object centering, which is inefficient. To overcome the above limitations, we propose LGCOAMix, an efficient context-aware and object-part-aware superpixel-based grid blending method for data augmentation. To the best of our knowledge, this is the first time that a label mixing strategy using a superpixel attention approach has been proposed for cutmix-based data augmentation. It is the first instance of learning local features from discriminative superpixel-wise regions and cross-image superpixel contrasts. Extensive experiments on various benchmark datasets show that LGCOAMix outperforms state-of-the-art cutmix-based data augmentation methods on classification tasks, {and weakly supervised object location on CUB200-2011.} We have demonstrated the effectiveness of LGCOAMix not only for CNN networks, but also for Transformer networks. Source codes are available at https://github.com/DanielaPlusPlus/LGCOAMix.

Paperid: 475, https://arxiv.org/pdf/2512.00086.pdf GitHub

Abstract:
Monocular depth estimation (MDE) plays a crucial role in enabling spatially-aware applications in Ultra-low-power (ULP) Internet-of-Things (IoT) platforms. However, the limited number of parameters of Deep Neural Networks for the MDE task, designed for IoT nodes, results in severe accuracy drops when the sensor data observed in the field shifts significantly from the training dataset. To address this domain shift problem, we present a multi-modal On-Device Learning (ODL) technique, deployed on an IoT device integrating a Greenwaves GAP9 MicroController Unit (MCU), a 80 mW monocular camera and a 8 x 8 pixel depth sensor, consuming $\approx$300mW. In its normal operation, this setup feeds a tiny 107 k-parameter $μ$PyD-Net model with monocular images for inference. The depth sensor, usually deactivated to minimize energy consumption, is only activated alongside the camera to collect pseudo-labels when the system is placed in a new environment. Then, the fine-tuning task is performed entirely on the MCU, using the new data. To optimize our backpropagation-based on-device training, we introduce a novel memory-driven sparse update scheme, which minimizes the fine-tuning memory to 1.2 MB, 2.2x less than a full update, while preserving accuracy (i.e., only 2% and 1.5% drops on the KITTI and NYUv2 datasets). Our in-field tests demonstrate, for the first time, that ODL for MDE can be performed in 17.8 minutes on the IoT node, reducing the root mean squared error from 4.9 to 0.6m with only 3 k self-labeled samples, collected in a real-life deployment scenario.

Paperid: 476, https://arxiv.org/pdf/2512.00085.pdf GitHub

Abstract:
Goal-conditioned policy learning for robotic manipulation presents significant challenges in maintaining performance across diverse objectives and environments. We introduce Hyper-GoalNet, a framework that generates task-specific policy network parameters from goal specifications using hypernetworks. Unlike conventional methods that simply condition fixed networks on goal-state pairs, our approach separates goal interpretation from state processing -- the former determines network parameters while the latter applies these parameters to current observations. To enhance representation quality for effective policy generation, we implement two complementary constraints on the latent space: (1) a forward dynamics model that promotes state transition predictability, and (2) a distance-based constraint ensuring monotonic progression toward goal states. We evaluate our method on a comprehensive suite of manipulation tasks with varying environmental randomization. Results demonstrate significant performance improvements over state-of-the-art methods, particularly in high-variability conditions. Real-world robotic experiments further validate our method's robustness to sensor noise and physical uncertainties. Code is available at: https://github.com/wantingyao/hyper-goalnet.

Paperid: 477, https://arxiv.org/pdf/2512.00077.pdf GitHub

Abstract:
The integration of Supernumerary Limbs (SLs) on humanoid robots poses a significant stability challenge due to the dynamic perturbations they introduce. This thesis addresses this issue by designing a novel hierarchical control architecture to improve humanoid locomotion stability with SLs. The core of this framework is a decoupled strategy that combines learning-based locomotion with model-based balancing. The low-level component consists of a walking gait for a Unitree H1 humanoid through imitation learning and curriculum learning. The high-level component actively utilizes the SLs for dynamic balancing. The effectiveness of the system is evaluated in a physics-based simulation under three conditions: baseline gait for an unladen humanoid (baseline walking), walking with a static SL payload (static payload), and walking with the active dynamic balancing controller (dynamic balancing). Our evaluation shows that the dynamic balancing controller improves stability. Compared to the static payload condition, the balancing strategy yields a gait pattern closer to the baseline and decreases the Dynamic Time Warping (DTW) distance of the CoM trajectory by 47\%. The balancing controller also improves the re-stabilization within gait cycles and achieves a more coordinated anti-phase pattern of Ground Reaction Forces (GRF). The results demonstrate that a decoupled, hierarchical design can effectively mitigate the internal dynamic disturbances arising from the mass and movement of the SLs, enabling stable locomotion for humanoids equipped with functional limbs. Code and videos are available here: https://github.com/heyzbw/HuSLs.

Paperid: 478, https://arxiv.org/pdf/2512.00074.pdf GitHub

Abstract:
Despite strong results on recognition and segmentation, current 3D visual pre-training methods often underperform on robotic manipulation. We attribute this gap to two factors: the lack of state-action-state dynamics modeling and the unnecessary redundancy of explicit geometric reconstruction. We introduce AFRO, a self-supervised framework that learns dynamics-aware 3D representations without action or reconstruction supervision. AFRO casts state prediction as a generative diffusion process and jointly models forward and inverse dynamics in a shared latent space to capture causal transition structure. To prevent feature leakage in action learning, we employ feature differencing and inverse-consistency supervision, improving the quality and stability of visual features. When combined with Diffusion Policy, AFRO substantially increases manipulation success rates across 16 simulated and 4 real-world tasks, outperforming existing pre-training approaches. The framework also scales favorably with data volume and task complexity. Qualitative visualizations indicate that AFRO learns semantically rich, discriminative features, offering an effective pre-training solution for 3D representation learning in robotics. Project page: https://kolakivy.github.io/AFRO/

Paperid: 479, https://arxiv.org/pdf/2512.00021.pdf GitHub

Abstract:
The emergence of multi-modal foundation models has markedly transformed the technology for autonomous driving, shifting away from conventional and mostly hand-crafted design choices towards unified, foundation-model-based approaches, capable of directly inferring motion trajectories from raw sensory inputs. This new class of methods can also incorporate natural language as an additional modality, with Vision-Language-Action (VLA) models serving as a representative example. In this review, we provide a comprehensive examination of such methods through a unifying taxonomy to critically evaluate their architectural design choices, methodological strengths, and their inherent capabilities and limitations. Our survey covers 37 recently proposed approaches that span the landscape of trajectory planning with foundation models. Furthermore, we assess these approaches with respect to the openness of their source code and datasets, offering valuable information to practitioners and researchers. We provide an accompanying webpage that catalogs the methods based on our taxonomy, available at: https://github.com/fiveai/FMs-for-driving-trajectories

Paperid: 480, https://arxiv.org/pdf/2512.00018.pdf GitHub

Abstract:
We present X-SYCON, a xylem-inspired multi-agent architecture in which coordination emerges from passive field dynamics rather than explicit planning or communication. Incidents (demands) and obstructions (hazards) continually write diffusing and decaying scalar fields, and agents greedily ascend a local utility $U=ϕ_{\mathrm{DE}}-κ\,ϕ_{\mathrm{HZ}}$ with light anti-congestion and separation. A beaconing rule triggered on first contact temporarily deepens the local demand sink, accelerating completion without reducing time-to-first-response. Across dynamic, partially blocked simulated environments, we observe low miss rates and stable throughput with interpretable, tunable trade-offs over carrier count, arrival rate, hazard density, and hazard sensitivity $κ$. We derive that a characteristic hydraulic length scale $\ell\approx\sqrt{D/λ}$ predicts recruitment range in a continuum approximation, and we provide a work-conservation (Ohm-law) bound consistent with sublinear capacity scaling with team size. Empirically: (i) soft hazard penalties yield fewer misses when obstacles already block motion; (ii) throughput saturates sublinearly with carriers while reliability improves sharply; (iii) stronger arrivals can reduce misses by sustaining sinks that recruit help; and (iv) phase-stability regions shrink with hazard density but are recovered by more carriers or higher arrivals. We refer to X-SYCON as an instance of Distributed Passive Computation and Control, and we evaluate it in simulations modeling communication-denied disaster response and other constrained sensing-action regimes.

Paperid: 481, https://arxiv.org/pdf/2512.00011.pdf GitHub

Abstract:
We present MRSeqStudio, a new all-in-one web-based tool for MRI sequence development and simulation, with the physics-based simulator KomaMRI running at the back-end and our own sequence designer at the front-end. It combines accessibility, interactivity and technical flexibility, within an environment suitable for both education and research. Our tool provides MR sequence design and simulation as a service, with no local installation needed by the user; alternatively, the code is publicly available on GitHub, for users who wish to deploy the application on their own server.

Paperid: 482, https://arxiv.org/pdf/2512.01797.pdf

Abstract:
Large language models (LLMs) frequently generate hallucinations -- plausible but factually incorrect outputs -- undermining their reliability. While prior work has examined hallucinations from macroscopic perspectives such as training data and objectives, the underlying neuron-level mechanisms remain largely unexplored. In this paper, we conduct a systematic investigation into hallucination-associated neurons (H-Neurons) in LLMs from three perspectives: identification, behavioral impact, and origins. Regarding their identification, we demonstrate that a remarkably sparse subset of neurons (less than $0.1\%$ of total neurons) can reliably predict hallucination occurrences, with strong generalization across diverse scenarios. In terms of behavioral impact, controlled interventions reveal that these neurons are causally linked to over-compliance behaviors. Concerning their origins, we trace these neurons back to the pre-trained base models and find that these neurons remain predictive for hallucination detection, indicating they emerge during pre-training. Our findings bridge macroscopic behavioral patterns with microscopic neural mechanisms, offering insights for developing more reliable LLMs.

Paperid: 483, https://arxiv.org/pdf/2512.06879.pdf

Abstract:
Researchers struggle to efficiently locate and manage relevant literature within the exponentially growing body of scientific publications. We present \textsc{WisPaper}, an intelligent academic retrieval and literature management platform that addresses this challenge through three integrated capabilities: (1) \textit{Scholar Search}, featuring both quick keyword-based and deep agentic search modes for efficient paper discovery; (2) \textit{Library}, a customizable knowledge base for systematic literature organization; and (3) \textit{AI Feeds}, an intelligent recommendation system that automatically delivers relevant new publications based on user interests. Unlike existing academic tools, \textsc{WisPaper} provides a closed-loop workflow that seamlessly connects literature discovery, management, and continuous tracking of research frontiers. Our multilingual and multidisciplinary system significantly reduces the time researchers from diverse backgrounds spend on paper screening and management, enabling them to focus on their core research activities. The platform is publicly accessible and serves researchers across academia and industry.

Paperid: 484, https://arxiv.org/pdf/2512.06726.pdf

Abstract:
Recent advances in fine-tuning multimodal large language models (MLLMs) using reinforcement learning have achieved remarkable progress, particularly with the introduction of various entropy control techniques. However, the role and characteristics of entropy in perception-oriented tasks like visual grounding, as well as effective strategies for controlling it, remain largely unexplored. To address this issue, we focus on the visual grounding task and analyze the role and characteristics of entropy in comparison to reasoning tasks. Building on these findings, we introduce ECVGPO (Entropy Control Visual Grounding Policy Optimization), an interpretable algorithm designed for effective entropy regulation. Through entropy control, the trade-off between exploration and exploitation is better balanced. Experiments show that ECVGPO achieves broad improvements across various benchmarks and models.

Paperid: 485, https://arxiv.org/pdf/2512.04987.pdf

Authors:AGI Team, Yuxuan Cai, Lu Chen, Qiaoling Chen, Yuyang Ding, Liwen Fan, Wenjie Fu, Yufei Gao, Honglin Guo, Pinxue Guo, Zhenhua Han, Zhengfu He, Hanglei Hu, Kai Hu, Shengjia Hua, Tianyu Huai, Baodai Huang, Li Ji, Zhen Jiang, Zhikai Lei, Bufan Li, Jiahang Lin, Lizhi Lin, Jinxiu Liu, Shichun Liu, Ziming Liu, Yuchen Ni, Pengfang Qian, Yujiong Shen, Qingyun Shi, Wentao Shu, Peng Sun, Yiran Suo, Tian Tang, Boyu Tian, Guoteng Wang, Junzhe Wang, Peixin Wang, Zhiheng Xi, Hang Yan, Jie Yang, Zhixiong Yang, Tianchu Yao, Guangze Ye, Qianxi Yu, Shuo Zhang, Xinyue Zhang, Yiqi Zhang, Jiarong Zhao, Miao Zheng, Rui Zheng, Enyu Zhou, Jiazheng Zhou, Maosen Zhou, Yuhao Zhou, Tao Gui, Yining Zheng, Xinchi Chen, Jie Zhou, Siyuan Feng, Qin Chen, Liang He, Qi Zhang, Xuanjing Huang, Xipeng Qiu

Abstract:
The evolution of Large Language Models (LLMs) from passive responders to autonomous agents necessitates a fundamental shift in learning paradigms -- from static imitation to incentive-driven decision making. However, this transition is significantly impeded by the lack of scalable infrastructure capable of constructing high-quality interaction signals for effective policy learning. To address this, we introduce a comprehensive method designed to systematically scale the diversity and complexity of interactive environments. Our method realizes this scaling by addressing three orthogonal dimensions: (1) Complexity: NexAU, a flexible agent framework that supports building complex agent hierarchies via simple configurations; (2) Diversity: NexA4A automatically generates diverse agent hierarchies from natural language to cover infinite domains; and (3) Fidelity: NexGAP bridges the simulation-reality gap by integrating dynamic real-world environment for grounded trajectories synthesis. We train Nex-N1 upon the diverse and complex interactive environments established by our infrastructure. Empirical results on benchmarks such as SWE-bench and tau2 demonstrate that Nex-N1 consistently outperforms SOTA open-source models and achieves competitive performance against frontier proprietary models on complex agentic tasks. We open-source the Nex ecosystem and model weights to facilitate further research.

Paperid: 486, https://arxiv.org/pdf/2512.04660.pdf

Abstract:
Image editing models are advancing rapidly, yet comprehensive evaluation remains a significant challenge. Existing image editing benchmarks generally suffer from limited task scopes, insufficient evaluation dimensions, and heavy reliance on manual annotations, which significantly constrain their scalability and practical applicability. To address this, we propose \textbf{I2I-Bench}, a comprehensive benchmark for image-to-image editing models, which features (i) diverse tasks, encompassing 10 task categories across both single-image and multi-image editing tasks, (ii) comprehensive evaluation dimensions, including 30 decoupled and fine-grained evaluation dimensions with automated hybrid evaluation methods that integrate specialized tools and large multimodal models (LMMs), and (iii) rigorous alignment validation, justifying the consistency between our benchmark evaluations and human preferences. Using I2I-Bench, we benchmark numerous mainstream image editing models, investigating the gaps and trade-offs between editing models across various dimensions. We will open-source all components of I2I-Bench to facilitate future research.

Paperid: 487, https://arxiv.org/pdf/2512.01340.pdf

Abstract:
Speech-driven Talking Human (TH) generation, commonly known as "Talker," currently faces limitations in multi-subject driving capabilities. Extending this paradigm to "Multi-Talker," capable of animating multiple subjects simultaneously, introduces richer interactivity and stronger immersion in audiovisual communication. However, current Multi-Talkers still exhibit noticeable quality degradation caused by technical limitations, resulting in suboptimal user experiences. To address this challenge, we construct THQA-MT, the first large-scale Multi-Talker-generated Talking Human Quality Assessment dataset, consisting of 5,492 Multi-Talker-generated THs (MTHs) from 15 representative Multi-Talkers using 400 real portraits collected online. Through subjective experiments, we analyze perceptual discrepancies among different Multi-Talkers and identify 12 common types of distortion. Furthermore, we introduce EvalTalker, a novel TH quality assessment framework. This framework possesses the ability to perceive global quality, human characteristics, and identity consistency, while integrating Qwen-Sync to perceive multimodal synchrony. Experimental results demonstrate that EvalTalker achieves superior correlation with subjective scores, providing a robust foundation for future research on high-quality Multi-Talker generation and evaluation.

Paperid: 488, https://arxiv.org/pdf/2512.00075.pdf

Abstract:
With the rapid progress in diffusion models, image synthesis has advanced to the stage of zero-shot image-to-image generation, where high-fidelity replication of facial identities or artistic styles can be achieved using just one portrait or artwork, without modifying any model weights. Although these techniques significantly enhance creative possibilities, they also pose substantial risks related to intellectual property violations, including unauthorized identity cloning and stylistic imitation. To counter such threats, this work presents Adapter Shield, the first universal and authentication-integrated solution aimed at defending personal images from misuse in zero-shot generation scenarios. We first investigate how current zero-shot methods employ image encoders to extract embeddings from input images, which are subsequently fed into the UNet of diffusion models through cross-attention layers. Inspired by this mechanism, we construct a reversible encryption system that maps original embeddings into distinct encrypted representations according to different secret keys. The authorized users can restore the authentic embeddings via a decryption module and the correct key, enabling normal usage for authorized generation tasks. For protection purposes, we design a multi-target adversarial perturbation method that actively shifts the original embeddings toward designated encrypted patterns. Consequently, protected images are embedded with a defensive layer that ensures unauthorized users can only produce distorted or encrypted outputs. Extensive evaluations demonstrate that our method surpasses existing state-of-the-art defenses in blocking unauthorized zero-shot image synthesis, while supporting flexible and secure access control for verified users.

Paperid: 489, https://arxiv.org/pdf/2512.01822.pdf

Abstract:
LLMs and Agents have achieved impressive progress in code generation, mathematical reasoning, and scientific discovery. However, existing benchmarks primarily measure correctness, overlooking the diversity of methods behind solutions. True innovation depends not only on producing correct answers but also on the originality of the approach. We present InnoGym, the first benchmark and framework designed to systematically evaluate the innovation potential of AI agents. InnoGym introduces two complementary metrics: performance gain, which measures improvement over the best-known solutions, and novelty, which captures methodological differences from prior approaches. The benchmark includes 18 carefully curated tasks from real-world engineering and scientific domains, each standardized through resource filtering, evaluator validation, and solution collection. In addition, we provide iGym, a unified execution environment for reproducible and long-horizon evaluations. Extensive experiments show that while some agents produce novel approaches, their lack of robustness limits performance gains. These results highlight a key gap between creativity and effectiveness, underscoring the need for benchmarks that evaluate both.

Paperid: 490, https://arxiv.org/pdf/2512.03722.pdf

Abstract:
Reinforcement Learning (RL) has shown remarkable success in enabling adaptive and data-driven optimization for various applications in wireless networks. However, classical RL suffers from limitations in generalization, learning feedback, interpretability, and sample efficiency in dynamic wireless environments. Large Language Models (LLMs) have emerged as a transformative Artificial Intelligence (AI) paradigm with exceptional capabilities in knowledge generalization, contextual reasoning, and interactive generation, which have demonstrated strong potential to enhance classical RL. This paper serves as a comprehensive tutorial on LLM-enhanced RL for wireless networks. We propose a taxonomy to categorize the roles of LLMs into four critical functions: state perceiver, reward designer, decision-maker, and generator. Then, we review existing studies exploring how each role of LLMs enhances different stages of the RL pipeline. Moreover, we provide a series of case studies to illustrate how to design and apply LLM-enhanced RL in low-altitude economy networking, vehicular networks, and space-air-ground integrated networks. Finally, we conclude with a discussion on potential future directions for LLM-enhanced RL and offer insights into its future development in wireless networks.

Paperid: 491, https://arxiv.org/pdf/2512.05774.pdf

Abstract:
Long video understanding (LVU) is challenging because answering real-world queries often depends on sparse, temporally dispersed cues buried in hours of mostly redundant and irrelevant content. While agentic pipelines improve video reasoning capabilities, prevailing frameworks rely on a query-agnostic captioner to perceive video information, which wastes computation on irrelevant content and blurs fine-grained temporal and spatial information. Motivated by active perception theory, we argue that LVU agents should actively decide what, when, and where to observe, and continuously assess whether the current observation is sufficient to answer the query. We present Active Video Perception (AVP), an evidence-seeking framework that treats the video as an interactive environment and acquires compact, queryrelevant evidence directly from pixels. Concretely, AVP runs an iterative plan-observe-reflect process with MLLM agents. In each round, a planner proposes targeted video interactions, an observer executes them to extract time-stamped evidence, and a reflector evaluates the sufficiency of the evidence for the query, either halting with an answer or triggering further observation. Across five LVU benchmarks, AVP achieves highest performance with significant improvements. Notably, AVP outperforms the best agentic method by 5.7% in average accuracy while only requires 18.4% inference time and 12.4% input tokens.

Paperid: 492, https://arxiv.org/pdf/2512.04545.pdf

Abstract:
Adjusting the outdated knowledge of large language models (LLMs) after deployment remains a major challenge. This difficulty has spurred the development of knowledge editing, which seeks to accurately and efficiently modify a model's internal (parametric) knowledge without retraining it from scratch. However, existing methods suffer from two limitations. First, they depend on structured triplets that are misaligned with the free-text nature of LLM pretraining and fail to capture the nuanced relationships among facts. Second, they typically support one-time knowledge updates, with relatively limited research on the problem of sequential or lifelong editing. To address these gaps, we propose a new task, Lifelong Free-text Knowledge Editing (LF-Edit), which enables models to incorporate updates expressed in natural language and supports continual editing over time. Despite its promise, LF-Edit faces the dual challenge of integrating new knowledge while mitigating the forgetting of prior information. To foster research on this new task, we construct a large-scale benchmark, Multi-Rank Lifelong Free-text Editing Benchmark (MRLF-Bench), containing 16,835 free-text edit requests. We further design a cognitively inspired multi-rank evaluation framework encompassing four levels: memorization, understanding, constrained comprehension, and reasoning. To tackle the challenges inherent in LF-Edit, we introduce a novel approach named EvoEdit that enhances knowledge injection through Latent Perturbation Augmentation and preserves prior information via Knowledge-driven Parameter Fusion. Experimental results demonstrate that EvoEdit substantially outperforms existing knowledge editing methods on the proposed LF-Edit task.

Paperid: 493, https://arxiv.org/pdf/2512.04324.pdf

Abstract:
Real-world enterprise data intelligence workflows encompass data engineering that turns raw sources into analytical-ready tables and data analysis that convert those tables into decision-oriented insights. We introduce DAComp, a benchmark of 210 tasks that mirrors these complex workflows. Data engineering (DE) tasks require repository-level engineering on industrial schemas, including designing and building multi-stage SQL pipelines from scratch and evolving existing systems under evolving requirements. Data analysis (DA) tasks pose open-ended business problems that demand strategic planning, exploratory analysis through iterative coding, interpretation of intermediate results, and the synthesis of actionable recommendations. Engineering tasks are scored through execution-based, multi-metric evaluation. Open-ended tasks are assessed by a reliable, experimentally validated LLM-judge, which is guided by hierarchical, meticulously crafted rubrics. Our experiments reveal that even state-of-the-art agents falter on DAComp. Performance on DE tasks is particularly low, with success rates under 20%, exposing a critical bottleneck in holistic pipeline orchestration, not merely code generation. Scores on DA tasks also average below 40%, highlighting profound deficiencies in open-ended reasoning and demonstrating that engineering and analysis are distinct capabilities. By clearly diagnosing these limitations, DAComp provides a rigorous and realistic testbed to drive the development of truly capable autonomous data agents for enterprise settings. Our data and code are available at https://da-comp.github.io

Paperid: 494, https://arxiv.org/pdf/2512.00971.pdf

Abstract:
The rapid advancement of humanoid robotics has intensified the need for robust and adaptable controllers to enable stable and efficient locomotion across diverse platforms. However, developing such controllers remains a significant challenge because existing solutions are tailored to specific robot designs, requiring extensive tuning of reward functions, physical parameters, and training hyperparameters for each embodiment. To address this challenge, we introduce H-Zero, a cross-humanoid locomotion pretraining pipeline that learns a generalizable humanoid base policy. We show that pretraining on a limited set of embodiments enables zero-shot and few-shot transfer to novel humanoid robots with minimal fine-tuning. Evaluations show that the pretrained policy maintains up to 81% of the full episode duration on unseen robots in simulation while enabling few-shot transfer to unseen humanoids and upright quadrupeds within 30 minutes of fine-tuning.

Paperid: 495, https://arxiv.org/pdf/2512.02874.pdf

Abstract:
Majority voting has proven effective for close-ended question answering by aggregating parallel reasoning traces. However, it is not directly applicable to open-ended reasoning, such as code generation and web-based deep research, where a "majority" over complete solutions is ill-defined. We introduce ThinkMerge, a training-free, plug-and-play decoding strategy that runs K parallel reasoning traces and averages their next-token logits at synchronization points to produce a single coherent output. ThinkMerge integrates seamlessly with vLLM/SGLang and remains compatible with standard decoding techniques such as Top-p/Top-k. Empirically, it matches or surpasses majority voting on AIME and GPQA, while delivering consistent gains on open-ended coding tasks: on LiveCodeBench (hard), pass@1 improves by +8.28% for DeepCoder-14B-Preview and +7.58% for Qwen3-8B. Beyond code, we further show that ThinkMerge improves web-based deep-research agents (e.g., WebSailor-7B/32B) across GAIA, BrowseComp-en/zh, and XbenchDeepSearch. These results demonstrate that parallel test-time scaling can benefit open-ended reasoning without relying on voting over complete outputs.

Paperid: 496, https://arxiv.org/pdf/2512.07464.pdf

Abstract:
For full-size humanoid robots, even with recent advances in reinforcement learning-based control, achieving reliable locomotion on complex terrains, such as long staircases, remains challenging. In such settings, limited perception, ambiguous terrain cues, and insufficient adaptation of gait timing can cause even a single misplaced or mistimed step to result in rapid loss of balance. We introduce a perceptive locomotion framework that merges terrain sensing, gait regulation, and whole-body control into a single reinforcement learning policy. A downward-facing depth camera mounted under the base observes the support region around the feet, and a compact U-Net reconstructs a dense egocentric height map from each frame in real time, operating at the same frequency as the control loop. The perceptual height map, together with proprioceptive observations, is processed by a unified policy that produces joint commands and a global stepping-phase signal, allowing gait timing and whole-body posture to be adapted jointly to the commanded motion and local terrain geometry. We further adopt a single-stage successive teacher-student training scheme for efficient policy learning and knowledge transfer. Experiments conducted on a 31-DoF, 1.65 m humanoid robot demonstrate robust locomotion in both simulation and real-world settings, including forward and backward stair ascent and descent, as well as crossing a 46 cm gap. Project Page:https://ga-phl.github.io/

Paperid: 497, https://arxiv.org/pdf/2512.00308.pdf

Abstract:
Dataset distillation seeks to synthesize a compact distilled dataset, enabling models trained on it to achieve performance comparable to models trained on the full dataset. Recent methods for large-scale datasets focus on matching global distributional statistics (e.g., mean and variance), but overlook critical instance-level characteristics and intraclass variations, leading to suboptimal generalization. We address this limitation by reformulating dataset distillation as an Optimal Transport (OT) distance minimization problem, enabling fine-grained alignment at both global and instance levels throughout the pipeline. OT offers a geometrically faithful framework for distribution matching. It effectively preserves local modes, intra-class patterns, and fine-grained variations that characterize the geometry of complex, high-dimensional distributions. Our method comprises three components tailored for preserving distributional geometry: (1) OT-guided diffusion sampling, which aligns latent distributions of real and distilled images; (2) label-image-aligned soft relabeling, which adapts label distributions based on the complexity of distilled image distributions; and (3) OT-based logit matching, which aligns the output of student models with soft-label distributions. Extensive experiments across diverse architectures and large-scale datasets demonstrate that our method consistently outperforms state-of-the-art approaches in an efficient manner, achieving at least 4% accuracy improvement under IPC=10 settings for each architecture on ImageNet-1K.

Paperid: 498, https://arxiv.org/pdf/2512.06192.pdf

Abstract:
Electronic devices are essential for robots but limit their usable environments. To overcome this, methods excluding electronics from the operating environment while retaining advanced electronic control and actuation have been explored. These include the remote hydraulic drive of electronics-free mobile robots, which offer high reachability, and long wire-driven robot arms with motors consolidated at the base, which offer high environmental resistance. To combine the advantages of both, this study proposes a new system, "Remote Wire Drive." As a proof-of-concept, we designed and developed the Remote Wire-Driven robot "REWW-ARM", which consists of the following components: 1) a novel power transmission mechanism, the "Remote Wire Transmission Mechanism" (RWTM), the key technology of the Remote Wire Drive; 2) an electronics-free distal mobile robot driven by it; and 3) a motor-unit that generates power and provides electronic closed-loop control based on state estimation via the RWTM. In this study, we evaluated the mechanical and control performance of REWW-ARM through several experiments, demonstrating its capability for locomotion, posture control, and object manipulation both on land and underwater. This suggests the potential for applying the Remote Wire-Driven system to various types of robots, thereby expanding their operational range.

Paperid: 499, https://arxiv.org/pdf/2512.03847.pdf

Abstract:
Reinforcement learning (RL) has shown strong performance in LLM post-training, but real-world deployment often involves noisy or incomplete supervision. In such settings, complex and unreliable supervision signals can destabilize training and harm generalization. While existing approaches such as worst-case optimization (e.g., RFQI, CQL) and mean-based methods (e.g., PPO, GRPO) can improve stability, they often overlook generalization and may produce overly conservative policies, leading to uneven performance across diverse real scenarios. To this end, we introduce DVPO (Distributional Value Modeling with Risk-aware Policy Optimization), a new RL framework that combines conditional risk theory with distributional value modeling to better balance robustness and generalization. DVPO learns token-level value distributions to provide fine-grained supervision, and applies an asymmetric risk regularization to shape the distribution tails: it contracts the lower tail to dampen noisy negative deviations, while expanding the upper tail to preserve exploratory diversity. Across extensive experiments and analysis in multi-turn dialogue, math reasoning, and scientific QA, DVPO consistently outperforms PPO, GRPO, and robust Bellman-based PPO under noisy supervision, showing its potential for LLM post-training in the real-world.

Paperid: 500, https://arxiv.org/pdf/2512.05111.pdf

Abstract:
Reward models are critical for aligning vision-language systems with human preferences, yet current approaches suffer from hallucination, weak visual grounding, and an inability to use tools for verification, limiting their reliability on complex multimodal reasoning tasks. We present ARM-Thinker, an A}gentic multimodal Reward Model that autonomously invokes external tools (e.g., image cropping, doc page retrieval) to ground judgments in verifiable evidence, replacing static, non-interactive reward scoring. This enables the model to verify fine-grained visual details, cross-reference multi-page evidence, and validate reasoning claims, which are capabilities absent in existing reward models. We train ARM-Thinker with multi-stage reinforcement learning, jointly optimizing tool-calling decisions and judgment accuracy. To evaluate agentic reward modeling, we introduce ARMBench-VL, comprising three benchmarks that assess fine-grained visual grounding (image-level tools), multi-page document understanding (retrieval tools), and instruction following (text-level verification). ARM-Thinker achieves +16.2% average improvement on reward modeling benchmarks, +9.6% on tool-use tasks, and outperforms baselines on multimodal math and logical reasoning benchmarks. Our results demonstrate that agentic capabilities significantly enhance both accuracy and interpretability of reward models.

Paperid: 501, https://arxiv.org/pdf/2512.06749.pdf

Abstract:
Large language model (LLM)-based multi-agent systems are challenging to debug because failures often arise from long, branching interaction traces. The prevailing practice is to leverage LLMs for log-based failure localization, attributing errors to a specific agent and step. However, this paradigm has two key limitations: (i) log-only debugging lacks validation, producing untested hypotheses, and (ii) single-step or single-agent attribution is often ill-posed, as we find that multiple distinct interventions can independently repair the failed task. To address the first limitation, we introduce DoVer, an intervention-driven debugging framework, which augments hypothesis generation with active verification through targeted interventions (e.g., editing messages, altering plans). For the second limitation, rather than evaluating on attribution accuracy, we focus on measuring whether the system resolves the failure or makes quantifiable progress toward task success, reflecting a more outcome-oriented view of debugging. Within the Magnetic-One agent framework, on the datasets derived from GAIA and AssistantBench, DoVer flips 18-28% of failed trials into successes, achieves up to 16% milestone progress, and validates or refutes 30-60% of failure hypotheses. DoVer also performs effectively on a different dataset (GSMPlus) and agent framework (AG2), where it recovers 49% of failed trials. These results highlight intervention as a practical mechanism for improving reliability in agentic systems and open opportunities for more robust, scalable debugging methods for LLM-based multi-agent systems. Project website and code will be available at https://aka.ms/DoVer.

Paperid: 502, https://arxiv.org/pdf/2512.00679.pdf

Abstract:
The powerful text understanding and generation capabilities of large language models (LLMs) have brought new vitality to general recommendation with implicit feedback. One possible strategy involves generating a unique user (or item) profile from historical interaction data, which is then mapped to a semantic representation in the language space. However, a single-instance profile may be insufficient to comprehensively capture the complex intentions behind a user's interacted items. Moreover, due to the inherent instability of LLMs, a biased or misinterpreted profile could even undermine the original recommendation performance. Consequently, an intuitive solution is to generate multiple profiles for each user (or item), each reflecting a distinct aspect of their characteristics. In light of this, we propose a unified recommendation framework with multi-faceted profile extrapolation (ProEx) in this paper. By leveraging chain-of-thought reasoning, we construct multiple distinct profiles for each user and item. These new profiles are subsequently mapped into semantic vectors, extrapolating from the position of the original profile to explore a broader region of the language space. Subsequently, we introduce the concept of environments, where each environment represents a possible linear combination of all profiles. The differences across environments are minimized to reveal the inherent invariance of user preferences. We apply ProEx to three discriminative methods and three generative methods, and conduct extensive experiments on three datasets. The experimental results demonstrate that ProEx significantly enhances the performance of these base recommendation models.

Paperid: 503, https://arxiv.org/pdf/2512.03556.pdf

Abstract:
Achieving generalizable embodied policies remains a key challenge. Traditional policy learning paradigms, including both Imitation Learning (IL) and Reinforcement Learning (RL), struggle to cultivate generalizability across diverse scenarios. While IL policies often overfit to specific expert trajectories, RL suffers from the inherent lack of a unified and general reward signal necessary for effective multi-scene generalization. We posit that the world model is uniquely capable of serving as a universal environment proxy to address this limitation. However, current world models primarily focus on their ability to predict observations and still rely on task-specific, handcrafted reward functions, thereby failing to provide a truly general training environment. Toward this problem, we propose RoboScape-R, a framework leveraging the world model to serve as a versatile, general-purpose proxy for the embodied environment within the RL paradigm. We introduce a novel world model-based general reward mechanism that generates ''endogenous'' rewards derived from the model's intrinsic understanding of real-world state transition dynamics. Extensive experiments demonstrate that RoboScape-R effectively addresses the limitations of traditional RL methods by providing an efficient and general training environment that substantially enhances the generalization capability of embodied policies. Our approach offers critical insights into utilizing the world model as an online training strategy and achieves an average 37.5% performance improvement over baselines under out-of-domain scenarios.

Paperid: 504, https://arxiv.org/pdf/2512.02472.pdf

Abstract:
AI self-evolution has long been envisioned as a path toward superintelligence, where models autonomously acquire, refine, and internalize knowledge from their own learning experiences. Yet in practice, unguided self-evolving systems often plateau quickly or even degrade as training progresses. These failures arise from issues such as concept drift, diversity collapse, and mis-evolution, as models reinforce their own biases and converge toward low-entropy behaviors. To enable models to self-evolve in a stable and controllable manner while minimizing reliance on human supervision, we introduce R-Few, a guided Self-Play Challenger-Solver framework that incorporates lightweight human oversight through in-context grounding and mixed training. At each iteration, the Challenger samples a small set of human-labeled examples to guide synthetic question generation, while the Solver jointly trains on human and synthetic examples under an online, difficulty-based curriculum. Across math and general reasoning benchmarks, R-Few achieves consistent and iterative improvements. For example, Qwen3-8B-Base improves by +3.0 points over R-Zero on math tasks and achieves performance on par with General-Reasoner, despite the latter being trained on 20 times more human data. Ablation studies confirm the complementary contributions of grounded challenger training and curriculum-based solver training, and further analysis shows that R-Few mitigates drift, yielding more stable and controllable co-evolutionary dynamics.

Paperid: 505, https://arxiv.org/pdf/2512.05485.pdf

Abstract:
While the deployment of large language models (LLMs) in high-value industries continues to expand, the systematic assessment of their safety against jailbreak and prompt-based attacks remains insufficient. Existing safety evaluation benchmarks and frameworks are often limited by an imbalanced integration of core components (attack, defense, and evaluation methods) and an isolation between flexible evaluation frameworks and standardized benchmarking capabilities. These limitations hinder reliable cross-study comparisons and create unnecessary overhead for comprehensive risk assessment. To address these gaps, we present TeleAI-Safety, a modular and reproducible framework coupled with a systematic benchmark for rigorous LLM safety evaluation. Our framework integrates a broad collection of 19 attack methods (including one self-developed method), 29 defense methods, and 19 evaluation methods (including one self-developed method). With a curated attack corpus of 342 samples spanning 12 distinct risk categories, the TeleAI-Safety benchmark conducts extensive evaluations across 14 target models. The results reveal systematic vulnerabilities and model-specific failure cases, highlighting critical trade-offs between safety and utility, and identifying potential defense patterns for future optimization. In practical scenarios, TeleAI-Safety can be flexibly adjusted with customized attack, defense, and evaluation combinations to meet specific demands. We release our complete code and evaluation results to facilitate reproducible research and establish unified safety baselines.

Paperid: 506, https://arxiv.org/pdf/2512.03422.pdf

Abstract:
In this paper, we provide a comprehensive overview of existing scene representation methods for robotics, covering traditional representations such as point clouds, voxels, signed distance functions (SDF), and scene graphs, as well as more recent neural representations like Neural Radiance Fields (NeRF), 3D Gaussian Splatting (3DGS), and the emerging Foundation Models. While current SLAM and localization systems predominantly rely on sparse representations like point clouds and voxels, dense scene representations are expected to play a critical role in downstream tasks such as navigation and obstacle avoidance. Moreover, neural representations such as NeRF, 3DGS, and foundation models are well-suited for integrating high-level semantic features and language-based priors, enabling more comprehensive 3D scene understanding and embodied intelligence. In this paper, we categorized the core modules of robotics into five parts (Perception, Mapping, Localization, Navigation, Manipulation). We start by presenting the standard formulation of different scene representation methods and comparing the advantages and disadvantages of scene representation across different modules. This survey is centered around the question: What is the best 3D scene representation for robotics? We then discuss the future development trends of 3D scene representations, with a particular focus on how the 3D Foundation Model could replace current methods as the unified solution for future robotic applications. The remaining challenges in fully realizing this model are also explored. We aim to offer a valuable resource for both newcomers and experienced researchers to explore the future of 3D scene representations and their application in robotics. We have published an open-source project on GitHub and will continue to add new works and technologies to this project.

Paperid: 507, https://arxiv.org/pdf/2512.00818.pdf

Abstract:
MLLMs MLLMs are beginning to appear in clinical workflows, but their ability to perform complex medical reasoning remains unclear. We present Med-CMR, a fine-grained Medical Complex Multimodal Reasoning benchmark. Med-CMR distinguishes from existing counterparts by three core features: 1) Systematic capability decomposition, splitting medical multimodal reasoning into fine-grained visual understanding and multi-step reasoning to enable targeted evaluation; 2) Challenging task design, with visual understanding across three key dimensions (small-object detection, fine-detail discrimination, spatial understanding) and reasoning covering four clinically relevant scenarios (temporal prediction, causal reasoning, long-tail generalization, multi-source integration); 3) Broad, high-quality data coverage, comprising 20,653 Visual Question Answering (VQA) pairs spanning 11 organ systems and 12 imaging modalities, validated via a rigorous two-stage (human expert + model-assisted) review to ensure clinical authenticity. We evaluate 18 state-of-the-art MLLMs with Med-CMR, revealing GPT-5 as the top-performing commercial model: 57.81 accuracy on multiple-choice questions (MCQs) and a 48.70 open-ended score, outperforming Gemini 2.5 Pro (49.87 MCQ accuracy, 45.98 open-ended score) and leading open-source model Qwen3-VL-235B-A22B (49.34 MCQ accuracy, 42.62 open-ended score). However, specialized medical MLLMs do not reliably outperform strong general models, and long-tail generalization emerges as the dominant failure mode. Med-CMR thus provides a stress test for visual-reasoning integration and rare-case robustness in medical MLLMs, and a rigorous yardstick for future clinical systems.

Paperid: 508, https://arxiv.org/pdf/2512.00300.pdf

Abstract:
Embodied 3D Semantic Scene Completion (SSC) infers dense geometry and semantics from continuous egocentric observations. Most existing Gaussian-based methods rely on random initialization of many primitives within predefined spatial bounds, resulting in redundancy and poor scalability to unbounded scenes. Recent depth-guided approach alleviates this issue but remains local, suffering from latency and memory overhead as scale increases. To overcome these challenges, we propose TGSFormer, a scalable Temporal Gaussian Splatting framework for embodied SSC. It maintains a persistent Gaussian memory for temporal prediction, without relying on image coherence or frame caches. For temporal fusion, a Dual Temporal Encoder jointly processes current and historical Gaussian features through confidence-aware cross-attention. Subsequently, a Confidence-aware Voxel Fusion module merges overlapping primitives into voxel-aligned representations, regulating density and maintaining compactness. Extensive experiments demonstrate that TGSFormer achieves state-of-the-art results on both local and embodied SSC benchmarks, offering superior accuracy and scalability with significantly fewer primitives while maintaining consistent long-term scene integrity. The code will be released upon acceptance.

Paperid: 509, https://arxiv.org/pdf/2512.03438.pdf

Abstract:
Agentic reasoning models trained with multimodal reinforcement learning (MMRL) have become increasingly capable, yet they are almost universally optimized using sparse, outcome-based rewards computed based on the final answers. Richer rewards computed from the reasoning tokens can improve learning significantly by providing more fine-grained guidance. However, it is challenging to compute more informative rewards in MMRL beyond those based on outcomes since different samples may require different scoring functions and teacher models may provide noisy reward signals too. In this paper, we introduce the Argos (Agentic Reward for Grounded & Objective Scoring), a principled reward agent to train multimodal reasoning models for agentic tasks. For each sample, Argos selects from a pool of teacher-model derived and rule-based scoring functions to simultaneously evaluate: (i) final response accuracy, (ii) spatiotemporal localization of referred entities and actions, and (iii) the quality of the reasoning process. We find that by leveraging our agentic verifier across both SFT data curation and RL training, our model achieves state-of-the-art results across multiple agentic tasks such as spatial reasoning, visual hallucination as well as robotics and embodied AI benchmarks. Critically, we demonstrate that just relying on SFT post-training on highly curated reasoning data is insufficient, as agents invariably collapse to ungrounded solutions during RL without our online verification. We also show that our agentic verifier can help to reduce reward-hacking in MMRL. Finally, we also provide a theoretical justification for the effectiveness of Argos through the concept of pareto-optimality.

Paperid: 510, https://arxiv.org/pdf/2512.02899.pdf

Abstract:
Diffusion models have achieved remarkable success in image generation, yet their deployment remains constrained by the heavy computational cost and the need for numerous inference steps. Previous efforts on fewer-step distillation attempt to skip redundant steps by training compact student models, yet they often suffer from heavy retraining costs and degraded generalization. In this work, we take a different perspective: we accelerate smartly, not evenly, applying smaller speedups to early semantic stages and larger ones to later redundant phases. We instantiate this phase-aware strategy with two experts that specialize in slow and fast denoising phases. Surprisingly, instead of investing massive effort in retraining student models, we find that simply equipping the base model with lightweight LoRA adapters achieves both efficient acceleration and strong generalization. We refer to these two adapters as Slow-LoRA and Fast-LoRA. Through extensive experiments, our method achieves up to 5 acceleration over the base model while maintaining comparable visual quality across diverse benchmarks. Remarkably, the LoRA experts are trained with only 1 samples on a single V100 within one hour, yet the resulting models generalize strongly on unseen prompts.

Paperid: 511, https://arxiv.org/pdf/2512.03870.pdf

Abstract:
Transformer decoders have achieved strong results across tasks, but the memory required for the KV cache becomes prohibitive at long sequence lengths. Although Cross-layer KV Cache sharing (e.g., YOCO, CLA) offers a path to mitigate KV Cache bottleneck, it typically underperforms within-layer methods like GQA. To understand the root cause, we investigate the information flow of keys and values of the top-layers. Our preliminary reveals a clear distribution: values are predominantly derived from the bottom layer, while keys draw more information from both bottom and middle layers. Building upon this, we propose FusedKV, whose top-layer KV caches are a learnable fusion of the most informative ones from the bottom and middle layers. This fusion operates directly on post-RoPE keys, preserving relative positional information without the computational cost of re-applying rotary embeddings. To further improve efficiency, we propose FusedKV-Lite, an cross-layer sharing approach, where top-layer KV caches are directly derived from the bottom-layer values and the middle-layer keys. Compared to FusedKV, FusedKV-Lite reduces I/O overhead at the cost of a slight increase in perplexity. In experiments on LLMs ranging from 332M to 4B parameters, our proposed method reduce 50\% cache memory while achieving lower validation perplexity than the standard Transformer decoder, establishing it as a memory-efficient, high-performance architectural alternative.

Paperid: 512, https://arxiv.org/pdf/2512.00552.pdf

Abstract:
Current evaluation of mathematical reasoning in language models relies primarily on answer accuracy, potentially masking fundamental failures in logical computation. We introduce a diagnostic framework that distinguishes genuine mathematical reasoning from superficial pattern matching through four complementary axes: forward-backward consistency, transitivity coverage, counterfactual sensitivity, and perturbation robustness. Through a case study applying this framework to Qwen3-0.6B on the MenatQA dataset, we reveal a striking disconnect between surface performance and reasoning fidelity. While the model achieves reasonable answer accuracy (70%+), it demonstrates poor backward consistency (15%), limited transitivity coverage (32.2%), and brittle sensitivity to perturbations. Our diagnostics expose reasoning failures invisible to traditional accuracy metrics, suggesting that this small model relies heavily on pattern matching rather than genuine logical computation. While our empirical findings are based on a single 600M-parameter model, the diagnostic framework itself is model-agnostic and generalizable. We release our evaluation protocols to enable the research community to assess reasoning fidelity across different model scales and architectures, moving beyond surface-level accuracy toward verifiable mathematical reasoning.

Paperid: 513, https://arxiv.org/pdf/2512.01382.pdf

Abstract:
Exemplar-guided Image Editing (EIE) aims to modify a source image according to a visual reference. Existing approaches often require large-scale pre-training to learn relationships between the source and reference images, incurring high computational costs. As a training-free alternative, inversion techniques can be used to map the source image into a latent space for manipulation. However, our empirical study reveals that standard inversion is sub-optimal for EIE, leading to poor quality and inefficiency. To tackle this challenge, we introduce \textbf{Reversible Inversion ({ReInversion})} for effective and efficient EIE. Specifically, ReInversion operates as a two-stage denoising process, which is first conditioned on the source image and subsequently on the reference. Besides, we introduce a Mask-Guided Selective Denoising (MSD) strategy to constrain edits to target regions, preserving the structural consistency of the background. Both qualitative and quantitative comparisons demonstrate that our ReInversion method achieves state-of-the-art EIE performance with the lowest computational overhead.

Paperid: 514, https://arxiv.org/pdf/2512.03044.pdf

Abstract:
Robust perception and dynamics modeling are fundamental to real-world robotic policy learning. Recent methods employ video diffusion models (VDMs) to enhance robotic policies, improving their understanding and modeling of the physical world. However, existing approaches overlook the coherent and physically consistent motion representations inherently encoded across frames in VDMs. To this end, we propose Video2Act, a framework that efficiently guides robotic action learning by explicitly integrating spatial and motion-aware representations. Building on the inherent representations of VDMs, we extract foreground boundaries and inter-frame motion variations while filtering out background noise and task-irrelevant biases. These refined representations are then used as additional conditioning inputs to a diffusion transformer (DiT) action head, enabling it to reason about what to manipulate and how to move. To mitigate inference inefficiency, we propose an asynchronous dual-system design, where the VDM functions as the slow System 2 and the DiT head as the fast System 1, working collaboratively to generate adaptive actions. By providing motion-aware conditions to System 1, Video2Act maintains stable manipulation even with low-frequency updates from the VDM. For evaluation, Video2Act surpasses previous state-of-the-art VLA methods by 7.7% in simulation and 21.7% in real-world tasks in terms of average success rate, further exhibiting strong generalization capabilities.

Paperid: 515, https://arxiv.org/pdf/2512.02013.pdf

Abstract:
Vision-Language-Action (VLA) models have recently emerged, demonstrating strong generalization in robotic scene understanding and manipulation. However, when confronted with long-horizon tasks that require defined goal states, such as LEGO assembly or object rearrangement, existing VLA models still face challenges in coordinating high-level planning with precise manipulation. Therefore, we aim to endow a VLA model with the capability to infer the "how" process from the "what" outcomes, transforming goal states into executable procedures. In this paper, we introduce ManualVLA, a unified VLA framework built upon a Mixture-of-Transformers (MoT) architecture, enabling coherent collaboration between multimodal manual generation and action execution. Unlike prior VLA models that directly map sensory inputs to actions, we first equip ManualVLA with a planning expert that generates intermediate manuals consisting of images, position prompts, and textual instructions. Building upon these multimodal manuals, we design a Manual Chain-of-Thought (ManualCoT) reasoning process that feeds them into the action expert, where each manual step provides explicit control conditions, while its latent representation offers implicit guidance for accurate manipulation. To alleviate the burden of data collection, we develop a high-fidelity digital-twin toolkit based on 3D Gaussian Splatting, which automatically generates manual data for planning expert training. ManualVLA demonstrates strong real-world performance, achieving an average success rate 32% higher than the previous hierarchical SOTA baseline on LEGO assembly and object rearrangement tasks.

Paperid: 516, https://arxiv.org/pdf/2512.03494.pdf

Abstract:
Large Language Models (LLMs) are increasingly prevalent in the field of long-context modeling, however, their inference computational costs have become a critical bottleneck hindering the advancement of tasks such as agents and multimodal applications. This report conducts a preliminary investigation into the effectiveness and theoretical mechanisms of the Top-$k$ Attention mechanism during both the decoding and training phases. First, we validate the effectiveness of exact Top-$k$ Decoding through extensive experimentation. Experiments demonstrate that retaining only the pivotal Keys with the highest similarity to the Query as the context window during the decoding stage achieves performance comparable to, or even surpassing, full attention on downstream tasks such as HELMET and LongBench v2. Second, we further explore the native Top-$k$ Attention training strategy. Experiments confirm that ensuring the consistency between training and inference regarding Top-$k$ Attention operations facilitates the further unlocking of Top-$k$ Decoding's potential, thereby significantly enhancing model performance. Furthermore, considering the high computational complexity of exact Top-$k$ Attention, we investigate the impact of approximate Top-$k$ algorithm precision on downstream tasks. Our research confirms a positive correlation between downstream task performance and approximation fidelity, and we provide statistical evaluations of the Lightning Indexer's precision within the DeepSeek-V3.2-Exp model. Finally, this report provides a theoretical interpretation from the perspective of Entropy. Experimental observations indicate that models subjected to Top-$k$ Attention SFT exhibit a distinct phenomenon of entropy reduction in downstream tasks, which validates the hypothesis that low-entropy states are better adapted to Top-$k$ Decoding.

Paperid: 517, https://arxiv.org/pdf/2512.07247.pdf

Abstract:
Recent studies have extended diffusion-based instruction-driven 2D image editing pipelines to 3D Gaussian Splatting (3DGS), enabling faithful manipulation of 3DGS assets and greatly advancing 3DGS content creation. However, it also exposes these assets to serious risks of unauthorized editing and malicious tampering. Although imperceptible adversarial perturbations against diffusion models have proven effective for protecting 2D images, applying them to 3DGS encounters two major challenges: view-generalizable protection and balancing invisibility with protection capability. In this work, we propose the first editing safeguard for 3DGS, termed AdLift, which prevents instruction-driven editing across arbitrary views and dimensions by lifting strictly bounded 2D adversarial perturbations into 3D Gaussian-represented safeguard. To ensure both adversarial perturbations effectiveness and invisibility, these safeguard Gaussians are progressively optimized across training views using a tailored Lifted PGD, which first conducts gradient truncation during back-propagation from the editing model at the rendered image and applies projected gradients to strictly constrain the image-level perturbation. Then, the resulting perturbation is backpropagated to the safeguard Gaussian parameters via an image-to-Gaussian fitting operation. We alternate between gradient truncation and image-to-Gaussian fitting, yielding consistent adversarial-based protection performance across different viewpoints and generalizes to novel views. Empirically, qualitative and quantitative results demonstrate that AdLift effectively protects against state-of-the-art instruction-driven 2D image and 3DGS editing.

Paperid: 518, https://arxiv.org/pdf/2512.07075.pdf

Abstract:
In recent years, large language models (LLMs) have demonstrated strong performance on multilingual tasks. Given its wide range of applications, cross-cultural understanding capability is a crucial competency. However, existing benchmarks for evaluating whether LLMs genuinely possess this capability suffer from three key limitations: a lack of contextual scenarios, insufficient cross-cultural concept mapping, and limited deep cultural reasoning capabilities. To address these gaps, we propose SAGE, a scenario-based benchmark built via cross-cultural core concept alignment and generative task design, to evaluate LLMs' cross-cultural understanding and reasoning. Grounded in cultural theory, we categorize cross-cultural capabilities into nine dimensions. Using this framework, we curated 210 core concepts and constructed 4530 test items across 15 specific real-world scenarios, organized under four broader categories of cross-cultural situations, following established item design principles. The SAGE dataset supports continuous expansion, and experiments confirm its transferability to other languages. It reveals model weaknesses across both dimensions and scenarios, exposing systematic limitations in cross-cultural reasoning. While progress has been made, LLMs are still some distance away from reaching a truly nuanced cross-cultural understanding. In compliance with the anonymity policy, we include data and code in the supplement materials. In future versions, we will make them publicly available online.

Paperid: 519, https://arxiv.org/pdf/2512.02473.pdf

Abstract:
Video world models have attracted significant attention for their ability to produce high-fidelity future visual observations conditioned on past observations and navigation actions. Temporally- and spatially-consistent, long-term world modeling has been a long-standing problem, unresolved with even recent state-of-the-art models, due to the prohibitively expensive computational costs for long-context inputs. In this paper, we propose WorldPack, a video world model with efficient compressed memory, which significantly improves spatial consistency, fidelity, and quality in long-term generation despite much shorter context length. Our compressed memory consists of trajectory packing and memory retrieval; trajectory packing realizes high context efficiency, and memory retrieval maintains the consistency in rollouts and helps long-term generations that require spatial reasoning. Our performance is evaluated with LoopNav, a benchmark on Minecraft, specialized for the evaluation of long-term consistency, and we verify that WorldPack notably outperforms strong state-of-the-art models.

Paperid: 520, https://arxiv.org/pdf/2512.01335.pdf

Abstract:
Retrieval-Augmented Generation (RAG) systems are increasingly central to robust AI, enhancing large language model (LLM) faithfulness by incorporating external knowledge. However, our study unveils a critical, overlooked vulnerability: their profound susceptibility to subtle symbolic perturbations, particularly through near-imperceptible emoticon tokens such as "(@_@)" that can catastrophically mislead retrieval, termed EmoRAG. We demonstrate that injecting a single emoticon into a query makes it nearly 100% likely to retrieve semantically unrelated texts that contain a matching emoticon. Our extensive experiment across general question-answering and code domains, using a range of state-of-the-art retrievers and generators, reveals three key findings: (I) Single-Emoticon Disaster: Minimal emoticon injections cause maximal disruptions, with a single emoticon almost 100% dominating RAG output. (II) Positional Sensitivity: Placing an emoticon at the beginning of a query can cause severe perturbation, with F1-Scores exceeding 0.92 across all datasets. (III) Parameter-Scale Vulnerability: Counterintuitively, models with larger parameters exhibit greater vulnerability to the interference. We provide an in-depth analysis to uncover the underlying mechanisms of these phenomena. Furthermore, we raise a critical concern regarding the robustness assumption of current RAG systems, envisioning a threat scenario where an adversary exploits this vulnerability to manipulate the RAG system. We evaluate standard defenses and find them insufficient against EmoRAG. To address this, we propose targeted defenses, analyzing their strengths and limitations in mitigating emoticon-based perturbations. Finally, we outline future directions for building robust RAG systems.

Paperid: 521, https://arxiv.org/pdf/2512.06493.pdf

Abstract:
The transition of cellular networks to (i) software-based systems on commodity hardware and (ii) platforms for services beyond connectivity introduces critical system-level challenges. As sensing emerges as a key feature toward 6G standardization, supporting Integrated Sensing and Communication (ISAC) with limited bandwidth and piggybacking on communication signals, while maintaining high reliability and performance, remains a fundamental challenge. In this paper, we provide two key contributions. First, we present a programmable, plug-and-play framework for processing PHY/MAC signals through real-time, GPU-accelerated Artificial Intelligence (AI) applications on the edge Radio Access Network (RAN) infrastructure. Building on the Open RAN dApp architecture, the framework interfaces with a GPU-accelerated gNB based on NVIDIA ARC-OTA, feeding PHY/MAC data to custom AI logic with latency under 0.5 ms for complex channel state information extraction. Second, we demonstrate the framework's capabilities through cuSense, an indoor localization dApp that consumes uplink DMRS channel estimates, removes static multipath components, and runs a neural network to infer the position of a moving person. Evaluated on a 3GPP-compliant 5G NR deployment, cuSense achieves a mean localization error of 77 cm, with 75% of predictions falling within 1 meter. This is without dedicated sensing hardware or modifications to the RAN stack or signals. We plan to release both the framework and cuSense pipelines as open source, providing a reference design for future AI-native RANs and ISAC applications.

Paperid: 522, https://arxiv.org/pdf/2512.06032.pdf

Abstract:
This paper investigates the fundamental discontinuity between the latest two Segment Anything Models: SAM2 and SAM3. We explain why the expertise in prompt-based segmentation of SAM2 does not transfer to the multimodal concept-driven paradigm of SAM3. SAM2 operates through spatial prompts points, boxes, and masks yielding purely geometric and temporal segmentation. In contrast, SAM3 introduces a unified vision-language architecture capable of open-vocabulary reasoning, semantic grounding, contrastive alignment, and exemplar-based concept understanding. We structure this analysis through five core components: (1) a Conceptual Break Between Prompt-Based and Concept-Based Segmentation, contrasting spatial prompt semantics of SAM2 with multimodal fusion and text-conditioned mask generation of SAM3; (2) Architectural Divergence, detailing pure vision-temporal design of SAM2 versus integration of vision-language encoders, geometry and exemplar encoders, fusion modules, DETR-style decoders, object queries, and ambiguity-handling via Mixture-of-Experts in SAM3; (3) Dataset and Annotation Differences, contrasting SA-V video masks with multimodal concept-annotated corpora of SAM3; (4) Training and Hyperparameter Distinctions, showing why SAM2 optimization knowledge does not apply to SAM3; and (5) Evaluation, Metrics, and Failure Modes, outlining the transition from geometric IoU metrics to semantic, open-vocabulary evaluation. Together, these analyses establish SAM3 as a new class of segmentation foundation model and chart future directions for the emerging concept-driven segmentation era.

Paperid: 523, https://arxiv.org/pdf/2512.07503.pdf

Abstract:
Large autoregressive models can generate high-quality, high-resolution images but suffer from slow generation speed, because these models require hundreds to thousands of sequential forward passes for next-token prediction during inference. To accelerate autoregressive text-to-image generation, we propose Speculative Jacobi Decoding++ (SJD++), a training-free probabilistic parallel decoding algorithm. Unlike traditional next-token prediction, SJD++ performs multi-token prediction in each forward pass, drastically reducing generation steps. Specifically, it integrates the iterative multi-token prediction mechanism from Jacobi decoding, with the probabilistic drafting-and-verification mechanism from speculative sampling. More importantly, for further acceleration, SJD++ reuses high-confidence draft tokens after each verification phase instead of resampling them all. We conduct extensive experiments on several representative autoregressive text-to-image generation models and demonstrate that SJD++ achieves $2\times$ to $3\times$ inference latency reduction and $2\times$ to $7\times$ step compression, while preserving visual quality with no observable degradation.

Paperid: 524, https://arxiv.org/pdf/2512.06690.pdf

Abstract:
Preference alignment has enabled large language models (LLMs) to better reflect human expectations, but current methods mostly optimize for population-level preferences, overlooking individual users. Personalization is essential, yet early approaches-such as prompt customization or fine-tuning-struggle to reason over implicit preferences, limiting real-world effectiveness. Recent "think-then-generate" methods address this by reasoning before response generation. However, they face challenges in long-form generation: their static one-shot reasoning must capture all relevant information for the full response generation, making learning difficult and limiting adaptability to evolving content. To address this issue, we propose FlyThinker, an efficient "think-while-generating" framework for personalized long-form generation. FlyThinker employs a separate reasoning model that generates latent token-level reasoning in parallel, which is fused into the generation model to dynamically guide response generation. This design enables reasoning and generation to run concurrently, ensuring inference efficiency. In addition, the reasoning model is designed to depend only on previous responses rather than its own prior outputs, which preserves training parallelism across different positions-allowing all reasoning tokens for training data to be produced in a single forward pass like standard LLM training, ensuring training efficiency. Extensive experiments on real-world benchmarks demonstrate that FlyThinker achieves better personalized generation while keeping training and inference efficiency.

Paperid: 525, https://arxiv.org/pdf/2512.04416.pdf

Abstract:
Data governance ensures data quality, security, and compliance through policies and standards, a critical foundation for scaling modern AI development. Recently, large language models (LLMs) have emerged as a promising solution for automating data governance by translating user intent into executable transformation code. However, existing benchmarks for automated data science often emphasize snippet-level coding or high-level analytics, failing to capture the unique challenge of data governance: ensuring the correctness and quality of the data itself. To bridge this gap, we introduce DataGovBench, a benchmark featuring 150 diverse tasks grounded in real-world scenarios, built on data from actual cases. DataGovBench employs a novel "reversed-objective" methodology to synthesize realistic noise and utilizes rigorous metrics to assess end-to-end pipeline reliability. Our analysis on DataGovBench reveals that current models struggle with complex, multi-step workflows and lack robust error-correction mechanisms. Consequently, we propose DataGovAgent, a framework utilizing a Planner-Executor-Evaluator architecture that integrates constraint-based planning, retrieval-augmented generation, and sandboxed feedback-driven debugging. Experimental results show that DataGovAgent significantly boosts the Average Task Score (ATS) on complex tasks from 39.7 to 54.9 and reduces debugging iterations by over 77.9 percent compared to general-purpose baselines.

Paperid: 526, https://arxiv.org/pdf/2512.00722.pdf

Abstract:
In this paper, we point out that the objective of the retrieval algorithms is to align with the LLM, which is similar to the objective of knowledge distillation in LLMs. We analyze the similarity in information focus between the distilled language model(DLM) and the original LLM from the perspective of information theory, and thus propose a novel paradigm that leverages a DLM as the retrieval algorithm. Based on the insight, we present SpeContext, an algorithm and system co-design for long-context reasoning. (1) At the algorithm level, SpeContext proposes lightweight retrieval head based on the head-level attention weights of DLM, achieving > 90% parameters reduction by pruning the redundancy. (2) At the system level, SpeContext designs an asynchronous prefetch dataflow via the elastic loading strategy, effectively overlapping KV cache retrieval with the LLM computation. (3) At the compilation level, SpeContext constructs the theoretical memory model and implements an adaptive memory management system to achieve acceleration by maximizing GPU memory utilization. We deploy and evaluate SpeContext in two resourceconstrained environments, cloud and edge. Extensive experiments show that, compared with the Huggingface framework, SpeContext achieves up to 24.89x throughput improvement in cloud and 10.06x speedup in edge with negligible accuracy loss, pushing the Pareto frontier of accuracy and throughput.

Paperid: 527, https://arxiv.org/pdf/2512.05591.pdf

Abstract:
Large language model post-training relies on reinforcement learning to improve model capability and alignment quality. However, the off-policy training paradigm introduces distribution shift, which often pushes the policy beyond the trust region, leading to training instabilities manifested as fluctuations in policy entropy and unstable gradients. Although PPO-Clip mitigates this issue through importance clipping, it still overlooks the global distributional shift of actions. To address these challenges, we propose using the entropy ratio between the current and previous policies as a new global metric that effectively quantifies the relative change in policy exploration throughout updates. Building on this metric, we introduce an \textbf{Entropy Ratio Clipping} (ERC) mechanism that imposes bidirectional constraints on the entropy ratio. This stabilizes policy updates at the global distribution level and compensates for the inability of PPO-clip to regulate probability shifts of un-sampled actions. We integrate ERC into both DAPO and GPPO reinforcement learning algorithms. Experiments across multiple benchmarks show that ERC consistently improves performance.

Paperid: 528, https://arxiv.org/pdf/2512.01939.pdf

Abstract:
The rise of large language models (LLMs) has sparked a surge of interest in agents, leading to the rapid growth of agent frameworks. Agent frameworks are software toolkits and libraries that provide standardized components, abstractions, and orchestration mechanisms to simplify agent development. Despite widespread use of agent frameworks, their practical applications and how they influence the agent development process remain underexplored. Different agent frameworks encounter similar problems during use, indicating that these recurring issues deserve greater attention and call for further improvements in agent framework design. Meanwhile, as the number of agent frameworks continues to grow and evolve, more than 80% of developers report difficulties in identifying the frameworks that best meet their specific development requirements. In this paper, we conduct the first empirical study of LLM-based agent frameworks, exploring real-world experiences of developers in building AI agents. To compare how well the agent frameworks meet developer needs, we further collect developer discussions for the ten previously identified agent frameworks, resulting in a total of 11,910 discussions. Finally, by analyzing these discussions, we compare the frameworks across five dimensions: development efficiency, functional abstraction, learning cost, performance optimization, and maintainability, which refers to how easily developers can update and extend both the framework itself and the agents built upon it over time. Our comparative analysis reveals significant differences among frameworks in how they meet the needs of agent developers. Overall, we provide a set of findings and implications for the LLM-driven AI agent framework ecosystem and offer insights for the design of future LLM-based agent frameworks and agent developers.

Paperid: 529, https://arxiv.org/pdf/2512.00380.pdf

Abstract:
In the context of software frameworks with limited resources (such as HarmonyOS), large language models (LLMs) often exhibit poor code generation performance because they lack sufficient exposure to such environments during pre-training. Although LLMs can usually maintain correct logical structures across programming languages, they frequently struggle when dealing with framework-specific APIs or syntax, resulting in errors. This indicates that while pre-training equips LLMs with general algorithmic capabilities, they remain unfamiliar with the distinctive syntax and API usage of underrepresented frameworks. As a result, even advanced commercial models like GPT-4o cannot reliably generate correct code without prior adaptation. To address this issue, we propose APIKG4SYN, a framework designed to exploit API knowledge graphs for the construction of API-oriented question-code pairs, specifically tailored for low-resource frameworks without requiring executable code. APIKG4SYN integrates both single-API and multi-API knowledge, where the latter is derived through uncertainty estimation (UE)-driven Monte Carlo Tree Search (MCTS), enabling the creation of a diverse and informative dataset for fine-tuning LLMs. Using HarmonyOS as a case study, we build the first benchmark for HarmonyOS code generation. Experimental results show that fine-tuning Qwen with APIKG4SYN raises pass@1 accuracy to 25.00%, compared with 17.59% for the baseline GPT model. These results confirm that API-oriented data significantly enhance LLM performance in low-resource software development scenarios.

Paperid: 530, https://arxiv.org/pdf/2512.07216.pdf

Abstract:
Lifelong user interest modeling is crucial for industrial recommender systems, yet existing approaches rely predominantly on ID-based features, suffering from poor generalization on long-tail items and limited semantic expressiveness. While recent work explores multimodal representations for behavior retrieval in the General Search Unit (GSU), they often neglect multimodal integration in the fine-grained modeling stage -- the Exact Search Unit (ESU). In this work, we present a systematic analysis of how to effectively leverage multimodal signals across both stages of the two-stage lifelong modeling framework. Our key insight is that simplicity suffices in the GSU: lightweight cosine similarity with high-quality multimodal embeddings outperforms complex retrieval mechanisms. In contrast, the ESU demands richer multimodal sequence modeling and effective ID-multimodal fusion to unlock its full potential. Guided by these principles, we propose MUSE, a simple yet effective multimodal search-based framework. MUSE has been deployed in Taobao display advertising system, enabling 100K-length user behavior sequence modeling and delivering significant gains in top-line metrics with negligible online latency overhead. To foster community research, we share industrial deployment practices and open-source the first large-scale dataset featuring ultra-long behavior sequences paired with high-quality multimodal embeddings. Our code and data is available at https://taobao-mm.github.io.

Paperid: 531, https://arxiv.org/pdf/2512.06862.pdf

Abstract:
In this paper, we propose a novel task termed Omni-Referring Image Segmentation (OmniRIS) towards highly generalized image segmentation. Compared with existing unimodally conditioned segmentation tasks, such as RIS and visual RIS, OmniRIS supports the input of text instructions and reference images with masks, boxes or scribbles as omni-prompts. This property makes it can well exploit the intrinsic merits of both text and visual modalities, i.e., granular attribute referring and uncommon object grounding, respectively. Besides, OmniRIS can also handle various segmentation settings, such as one v.s. many and many v.s. many, further facilitating its practical use. To promote the research of OmniRIS, we also rigorously design and construct a large dataset termed OmniRef, which consists of 186,939 omni-prompts for 30,956 images, and establish a comprehensive evaluation system. Moreover, a strong and general baseline termed OmniSegNet is also proposed to tackle the key challenges of OmniRIS, such as omni-prompt encoding. The extensive experiments not only validate the capability of OmniSegNet in following omni-modal instructions, but also show the superiority of OmniRIS for highly generalized image segmentation.

Paperid: 532, https://arxiv.org/pdf/2512.03025.pdf

Abstract:
Achievement. We introduce LORE, a systematic framework for Large Generative Model-based relevance in e-commerce search. Deployed and iterated over three years, LORE achieves a cumulative +27\% improvement in online GoodRate metrics. This report shares the valuable experience gained throughout its development lifecycle, spanning data, features, training, evaluation, and deployment. Insight. While existing works apply Chain-of-Thought (CoT) to enhance relevance, they often hit a performance ceiling. We argue this stems from treating relevance as a monolithic task, lacking principled deconstruction. Our key insight is that relevance comprises distinct capabilities: knowledge and reasoning, multi-modal matching, and rule adherence. We contend that a qualitative-driven decomposition is essential for breaking through current performance bottlenecks. Contributions. LORE provides a complete blueprint for the LLM relevance lifecycle. Key contributions include: (1) A two-stage training paradigm combining progressive CoT synthesis via SFT with human preference alignment via RL. (2) A comprehensive benchmark, RAIR, designed to evaluate these core capabilities. (3) A query frequency-stratified deployment strategy that efficiently transfers offline LLM capabilities to the online system. LORE serves as both a practical solution and a methodological reference for other vertical domains.

Paperid: 533, https://arxiv.org/pdf/2512.01214.pdf

Abstract:
In the contemporary digital landscape, multi-modal media manipulation has emerged as a significant societal threat, impacting the reliability and integrity of information dissemination. Current detection methodologies in this domain often overlook the crucial aspect of localized information, despite the fact that manipulations frequently occur in specific areas, particularly in facial regions. In response to this critical observation, we propose the M4-BLIP framework. This innovative framework utilizes the BLIP-2 model, renowned for its ability to extract local features, as the cornerstone for feature extraction. Complementing this, we incorporate local facial information as prior knowledge. A specially designed alignment and fusion module within M4-BLIP meticulously integrates these local and global features, creating a harmonious blend that enhances detection accuracy. Furthermore, our approach seamlessly integrates with Large Language Models (LLM), significantly improving the interpretability of the detection outcomes. Extensive quantitative and visualization experiments validate the effectiveness of our framework against the state-of-the-art competitors.

Paperid: 534, https://arxiv.org/pdf/2512.00961.pdf

Abstract:
Reinforcement Learning (RL) has achieved remarkable success in various domains, yet it often relies on carefully designed programmatic reward functions to guide agent behavior. Designing such reward functions can be challenging and may not generalize well across different tasks. To address this limitation, we leverage the rich world knowledge contained in pretrained video diffusion models to provide goal-driven reward signals for RL agents without ad-hoc design of reward. Our key idea is to exploit off-the-shelf video diffusion models pretrained on large-scale video datasets as informative reward functions in terms of video-level and frame-level goals. For video-level rewards, we first finetune a pretrained video diffusion model on domain-specific datasets and then employ its video encoder to evaluate the alignment between the latent representations of agent's trajectories and the generated goal videos. To enable more fine-grained goal-achievement, we derive a frame-level goal by identifying the most relevant frame from the generated video using CLIP, which serves as the goal state. We then employ a learned forward-backward representation that represents the probability of visiting the goal state from a given state-action pair as frame-level reward, promoting more coherent and goal-driven trajectories. Experiments on various Meta-World tasks demonstrate the effectiveness of our approach.

Paperid: 535, https://arxiv.org/pdf/2512.00076.pdf

Abstract:
We contend that embodied learning is fundamentally a lifecycle problem rather than a single-stage optimization. Systems that optimize only one link (data collection, simulation, learning, or deployment) rarely sustain improvement or generalize beyond narrow settings. We introduce Arcadia, a closed-loop framework that operationalizes embodied lifelong learning by tightly coupling four stages: (1) Self-evolving exploration and grounding for autonomous data acquisition in physical environments, (2) Generative scene reconstruction and augmentation for realistic and extensible scene creation, (3) a Shared embodied representation architecture that unifies navigation and manipulation within a single multimodal backbone, and (4) Sim-from-real evaluation and evolution that closes the feedback loop through simulation-based adaptation. This coupling is non-decomposable: removing any stage breaks the improvement loop and reverts to one-shot training. Arcadia delivers consistent gains on navigation and manipulation benchmarks and transfers robustly to physical robots, indicating that a tightly coupled lifecycle: continuous real-world data acquisition, generative simulation update, and shared-representation learning, supports lifelong improvement and end-to-end generalization. We release standardized interfaces enabling reproducible evaluation and cross-model comparison in reusable environments, positioning Arcadia as a scalable foundation for general-purpose embodied agents.

Paperid: 536, https://arxiv.org/pdf/2512.04123.pdf

Abstract:
AI agents are actively running in production across diverse industries, yet little is publicly known about which technical approaches enable successful real-world deployments. We present the first large-scale systematic study of AI agents in production, surveying 306 practitioners and conducting 20 in-depth case studies via interviews across 26 domains. We investigate why organizations build agents, how they build them, how they evaluate them, and what the top development challenges are. We find that production agents are typically built using simple, controllable approaches: 68% execute at most 10 steps before requiring human intervention, 70% rely on prompting off-the-shelf models instead of weight tuning, and 74% depend primarily on human evaluation. Reliability remains the top development challenge, driven by difficulties in ensuring and evaluating agent correctness. Despite these challenges, simple yet effective methods already enable agents to deliver impact across diverse industries. Our study documents the current state of practice and bridges the gap between research and deployment by providing researchers visibility into production challenges while offering practitioners proven patterns from successful deployments.

Paperid: 537, https://arxiv.org/pdf/2512.01707.pdf

Abstract:
Streaming video understanding requires models not only to process temporally incoming frames, but also to anticipate user intention for realistic applications like AR glasses. While prior streaming benchmarks evaluate temporal reasoning, none measure whether MLLMs can interpret or leverage human gaze signals within a streaming setting. To fill this gap, we introduce StreamGaze, the first benchmark designed to evaluate how effectively MLLMs use gaze for temporal and proactive reasoning in streaming videos. StreamGaze introduces gaze-guided past, present, and proactive tasks that comprehensively evaluate streaming video understanding. These tasks assess whether models can use real-time gaze to follow shifting attention and infer user intentions from only past and currently observed frames. To build StreamGaze, we develop a gaze-video QA generation pipeline that aligns egocentric videos with raw gaze trajectories via fixation extraction, region-specific visual prompting, and scanpath construction. This pipeline produces spatio-temporally grounded QA pairs that closely reflect human perceptual dynamics. Across all StreamGaze tasks, we observe substantial performance gaps between state-of-the-art MLLMs and human performance, revealing fundamental limitations in gaze-based temporal reasoning, intention modeling, and proactive prediction. We further provide detailed analyses of gaze-prompting strategies, reasoning behaviors, and task-specific failure modes, offering deeper insight into why current MLLMs struggle and what capabilities future models must develop. All data and code will be publicly released to support continued research in gaze-guided streaming video understanding.

Paperid: 538, https://arxiv.org/pdf/2512.05714.pdf

Abstract:
Over the past four decades, artificial intelligence (AI) research has flourished at the nexus of academia and industry. However, Big Tech companies have increasingly acquired the edge in computational resources, big data, and talent. So far, it has been largely unclear how many papers the industry funds, how their citation impact compares to non-funded papers, and what drives industry interest. This study fills that gap by quantifying the number of industry-funded papers at 10 top AI conferences (e.g., ICLR, CVPR, AAAI, ACL) and their citation influence. We analyze about 49.8K papers, about 1.8M citations from AI papers to other papers, and about 2.3M citations from other papers to AI papers from 1998-2022 in Scopus. Through seven research questions, we examine the volume and evolution of industry funding in AI research, the citation impact of funded papers, the diversity and temporal range of their citations, and the subfields in which industry predominantly acts. Our findings reveal that industry presence has grown markedly since 2015, from less than 2 percent to more than 11 percent in 2020. Between 2018 and 2022, 12 percent of industry-funded papers achieved high citation rates as measured by the h5-index, compared to 4 percent of non-industry-funded papers and 2 percent of non-funded papers. Top AI conferences engage more with industry-funded research than non-funded research, as measured by our newly proposed metric, the Citation Preference Ratio (CPR). We show that industry-funded research is increasingly insular, citing predominantly other industry-funded papers while referencing fewer non-funded papers. These findings reveal new trends in AI research funding, including a shift towards more industry-funded papers and their growing citation impact, greater insularity of industry-funded work than non-funded work, and a preference of industry-funded research to cite recent work.

Paperid: 539, https://arxiv.org/pdf/2512.04501.pdf

Abstract:
Generative models have shown immense potential for wireless communication by learning complex channel data distributions. However, the iterative denoising process associated with these models imposes a significant challenge in latency-sensitive wireless communication scenarios, particularly in channel estimation. To address this challenge, we propose a novel solution for one-step generative channel estimation. Our approach bypasses the time-consuming iterative steps of conventional models by directly learning the average velocity field. Through extensive simulations, we validate the effectiveness of our proposed method over existing state-of-the-art diffusion-based approach. Specifically, our scheme achieves a normalized mean squared error up to 2.65 dB lower than the diffusion method and reduces latency by around 90%, demonstrating the potential of our method to enhance channel estimation performance.

Paperid: 540, https://arxiv.org/pdf/2512.03067.pdf

Abstract:
Nowadays, recommendation systems have become crucial to online platforms, shaping user exposure by accurate preference modeling. However, such an exposure strategy can also reinforce users' existing preferences, leading to a notorious phenomenon named filter bubbles. Given its negative effects, such as group polarization, increasing attention has been paid to exploring reasonable measures to filter bubbles. However, most existing evaluation metrics simply measure the diversity of user exposure, failing to distinguish between algorithmic preference modeling and actual information confinement. In view of this, we introduce Bubble Escape Potential (BEP), a behavior-aware measure that quantifies how easily users can escape from filter bubbles. Specifically, BEP leverages a contrastive simulation framework that assigns different behavioral tendencies (e.g., positive vs. negative) to synthetic users and compares the induced exposure patterns. This design enables decoupling the effect of filter bubbles and preference modeling, allowing for more precise diagnosis of bubble severity. We conduct extensive experiments across multiple recommendation models to examine the relationship between predictive accuracy and bubble escape potential across different groups. To the best of our knowledge, our empirical results are the first to quantitatively validate the dilemma between preference modeling and filter bubbles. What's more, we observe a counter-intuitive phenomenon that mild random recommendations are ineffective in alleviating filter bubbles, which can offer a principled foundation for further work in this direction.

Paperid: 541, https://arxiv.org/pdf/2512.01213.pdf

Abstract:
As a variant of the Area Under the ROC Curve (AUC), the partial AUC (PAUC) focuses on a specific range of false positive rate (FPR) and/or true positive rate (TPR) in the ROC curve. It is a pivotal evaluation metric in real-world scenarios with both class imbalance and decision constraints. However, selecting instances within these constrained intervals during its calculation is NP-hard, and thus typically requires approximation techniques for practical resolution. Despite the progress made in PAUC optimization over the last few years, most existing methods still suffer from uncontrollable approximation errors or a limited scalability when optimizing the approximate PAUC objectives. In this paper, we close the approximation gap of PAUC optimization by presenting two simple instance-wise minimax reformulations: one with an asymptotically vanishing gap, the other with the unbiasedness at the cost of more variables. Our key idea is to first establish an equivalent instance-wise problem to lower the time complexity, simplify the complicated sample selection procedure by threshold learning, and then apply different smoothing techniques. Equipped with an efficient solver, the resulting algorithms enjoy a linear per-iteration computational complexity w.r.t. the sample size and a convergence rate of $O(ε^{-1/3})$ for typical one-way and two-way PAUCs. Moreover, we provide a tight generalization bound of our minimax reformulations. The result explicitly demonstrates the impact of the TPR/FPR constraints $α$/$β$ on the generalization and exhibits a sharp order of $\tilde{O}(α^{-1}\n_+^{-1} + β^{-1}\n_-^{-1})$. Finally, extensive experiments on several benchmark datasets validate the strength of our proposed methods.

Paperid: 542, https://arxiv.org/pdf/2512.06915.pdf

Abstract:
Automated environment configuration is a critical bottleneck in scaling software engineering (SWE) automation. To provide a reliable evaluation standard for this task, we present Multi-Docker-Eval benchmark. It includes 40 real-world repositories spanning 9 programming languages and measures both success in achieving executable states and efficiency under realistic constraints. Our extensive evaluation of state-of-the-art LLMs and agent frameworks reveals key insights: (1) the overall success rate of current models is low (F2P at most 37.7%), with environment construction being the primary bottleneck; (2) model size and reasoning length are not decisive factors, and open-source models like DeepSeek-V3.1 and Kimi-K2 are competitive in both efficiency and effectiveness; (3) agent framework and programming language also have significantly influence on success rate. These findings provide actionable guidelines for building scalable, fully automated SWE pipelines.

Paperid: 543, https://arxiv.org/pdf/2512.04532.pdf

Abstract:
Video Large Language Models (Video LLMs) have shown impressive performance across a wide range of video-language tasks. However, they often fail in scenarios requiring a deeper understanding of physical dynamics. This limitation primarily arises from their reliance on appearance-based matching. Incorporating physical motion modeling is crucial for deeper video understanding, but presents three key challenges: (1) motion signals are often entangled with appearance variations, making it difficult to extract clean physical cues; (2) effective motion modeling requires not only continuous-time motion representations but also capturing physical dynamics; and (3) collecting accurate annotations for physical attributes is costly and often impractical. To address these issues, we propose PhyVLLM, a physical-guided video-language framework that explicitly incorporates physical motion into Video LLMs. Specifically, PhyVLLM disentangles visual appearance and object motion through a dual-branch encoder. To model physical dynamics over time, we incorporate a Neural Ordinary Differential Equation (Neural ODE) module, which generates differentiable physical dynamic representations. The resulting motion-aware representations are projected into the token space of a pretrained LLM, enabling physics reasoning without compromising the model's original multimodal capabilities. To circumvent the need for explicit physical labels, PhyVLLM employs a self-supervised manner to model the continuous evolution of object motion. Experimental results demonstrate that PhyVLLM significantly outperforms state-of-the-art Video LLMs on both physical reasoning and general video understanding tasks, highlighting the advantages of incorporating explicit physical modeling.

Paperid: 544, https://arxiv.org/pdf/2512.04513.pdf

Abstract:
Building generalist embodied agents requires a unified system that can interpret multimodal goals, model environment dynamics, and execute reliable actions across diverse real-world tasks. Multimodal large language models (MLLMs) offer strong semantic priors and cross-modal generalization, while world models (WMs) provide actionable latent dynamics for prediction and control. Their combination holds promise for open-ended embodied intelligence, yet introduces two key challenges: (1) establishing a tight coupling between the semantic intent from MLLMs and the dynamic state representations within the WM's latent space, and (2) achieving task-aware adaptability that supports multi-task learning and cross-environment generalization. To address these limitations, we propose BiTAgent, a task-aware dynamic joint framework that enables bidirectional coupling between MLLMs and WMs. BiTAgent establishes two complementary pathways: a forward path that injects MLLM representations into the WM's latent space for semantically guided imagination, and a backward path where WM-generated feedback refines the MLLM's semantic space via dense text-conditioned rewards. This bidirectional interaction is realized through three synergistic components: Task-Aware Dynamic Joint Learning, Task-Aware Behavior Learning, and MLLM-WM Joint Optimization, which together harmonize semantic reasoning and dynamic prediction. Extensive experiments across multi-task and cross-environment settings demonstrate superior stability and generalization over state-of-the-art baselines, marking a step toward open-ended embodied learning.

Paperid: 545, https://arxiv.org/pdf/2512.01342.pdf

Abstract:
Large-scale video-text pretraining achieves strong performance but depends on noisy, synthetic captions with limited semantic coverage, often overlooking implicit world knowledge such as object motion, 3D geometry, and physical cues. In contrast, masked video modeling (MVM) directly exploits spatiotemporal structures but trails text-supervised methods on general tasks. We find this gap arises from overlooked architectural issues: pixel-level reconstruction struggles with convergence and its low-level requirement often conflicts with semantics, while latent prediction often encourages shortcut learning. To address these, we disentangle the traditional encoder-decoder design into an Encoder-Predictor-Decoder (EPD) framework, where the predictor acts as a latent world model, and propose InternVideo-Next, a two-stage pretraining scheme that builds a semantically consistent yet detail-preserving latent space for this world model. First, conventional linear decoder in pixel MVM enforces the predictor output latent to be linearly projected to, thus separable in pixel space, causing the conflict with semantic abstraction. Our Stage 1 proposes a conditional diffusion decoder and injects reliable image-level semantic priors to enhance semantics and convergence, thus bridging pixel-level fidelity with high-level semantic abstraction. Stage 2 further learns world knowledge by predicting frozen Stage 1 targets within this space, mitigating shortcut learning. Trained on public, unlabeled videos, InternVideo-Next achieves state-of-the-art results across benchmarks and provides a scalable path toward general video representation learning.

Paperid: 546, https://arxiv.org/pdf/2512.00711.pdf

Abstract:
Semantic communication can significantly improve bandwidth utilization in wireless systems by exploiting the meaning behind raw data. However, the advancements achieved through semantic communication are closely dependent on the development of deep learning (DL) models for joint source-channel coding (JSCC) encoder/decoder techniques, which require a large amount of data for training. To address this data-intensive nature of DL models, federated learning (FL) has been proposed to train a model in a distributed manner, where the server broadcasts the DL model to clients in the network for training with their local data. However, the conventional FL approaches suffer from catastrophic degradation when client data are from different domains. In contrast, in this paper, a novel FL framework is proposed to address this domain shift by constructing the global representation, which aligns with the local features of the clients to preserve the semantics of different data domains. In addition, the dominance problem of client domains with a large number of samples is identified and, then, addressed with a domain-aware aggregation approach. This work is the first to consider the domain shift in training the semantic communication system for the image reconstruction task. Finally, simulation results demonstrate that the proposed approach outperforms the model-contrastive FL (MOON) framework by 0.5 for PSNR values under three domains at an SNR of 1 dB, and this gap continues to widen as the channel quality improves.

Paperid: 547, https://arxiv.org/pdf/2512.02471.pdf

Abstract:
Cell clustering is crucial for uncovering cellular heterogeneity in single-cell RNA sequencing (scRNA-seq) data by identifying cell types and marker genes. Despite its importance, benchmarks for scRNA-seq clustering methods remain fragmented, often lacking standardized protocols and failing to incorporate recent advances in artificial intelligence. To fill these gaps, we present scCluBench, a comprehensive benchmark of clustering algorithms for scRNA-seq data. First, scCluBench provides 36 scRNA-seq datasets collected from diverse public sources, covering multiple tissues, which are uniformly processed and standardized to ensure consistency for systematic evaluation and downstream analyses. To evaluate performance, we collect and reproduce a range of scRNA-seq clustering methods, including traditional, deep learning-based, graph-based, and biological foundation models. We comprehensively evaluate each method both quantitatively and qualitatively, using core performance metrics as well as visualization analyses. Furthermore, we construct representative downstream biological tasks, such as marker gene identification and cell type annotation, to further assess the practical utility. scCluBench then investigates the performance differences and applicability boundaries of various clustering models across diverse analytical tasks, systematically assessing their robustness and scalability in real-world scenarios. Overall, scCluBench offers a standardized and user-friendly benchmark for scRNA-seq clustering, with curated datasets, unified evaluation protocols, and transparent analyses, facilitating informed method selection and providing valuable insights into model generalizability and application scope.

Paperid: 548, https://arxiv.org/pdf/2512.01725.pdf

Abstract:
Large Language Models (LLMs) excel in reasoning tasks requiring a single correct answer, but they perform poorly in multi-solution tasks that require generating comprehensive and diverse answers. We attribute this limitation to \textbf{reasoning overconfidence}: a tendency to express undue certainty in an incomplete solution set. To examine the effect, we introduce \textit{MuSoBench}, a benchmark of multi-solution problems. Experiments show that the conventional short chain-of-thought (Short-CoT) prompting paradigm exhibits pronounced overconfidence, whereas the emerging long chain-of-thought (Long-CoT) approach mitigates it through iterative exploration and self-reflection. We further characterise observable behaviours and influential factors. To probe the underlying cause, we propose the \textbf{cognitive-rigidity hypothesis}, which posits that overconfidence arises when the reasoning process prematurely converges on a narrow set of thought paths. An attention-entropy analysis offers preliminary support for this view. These findings provide tools for assessing the completeness of LLM reasoning and highlight the need to move evaluation beyond single-answer accuracy toward comprehensive exploration.

Paperid: 549, https://arxiv.org/pdf/2512.04385.pdf

Abstract:
Fine-grained air pollution forecasting is crucial for urban management and the development of healthy buildings. Deploying portable sensors on mobile platforms such as cars and buses offers a low-cost, easy-to-maintain, and wide-coverage data collection solution. However, due to the random and uncontrollable movement patterns of these non-dedicated mobile platforms, the resulting sensor data are often incomplete and temporally inconsistent. By exploring potential training patterns in the reverse process of diffusion models, we propose Spatio-Temporal Physics-Informed Diffusion Models (STeP-Diff). STeP-Diff leverages DeepONet to model the spatial sequence of measurements along with a PDE-informed diffusion model to forecast the spatio-temporal field from incomplete and time-varying data. Through a PDE-constrained regularization framework, the denoising process asymptotically converges to the convection-diffusion dynamics, ensuring that predictions are both grounded in real-world measurements and aligned with the fundamental physics governing pollution dispersion. To assess the performance of the system, we deployed 59 self-designed portable sensing devices in two cities, operating for 14 days to collect air pollution data. Compared to the second-best performing algorithm, our model achieved improvements of up to 89.12% in MAE, 82.30% in RMSE, and 25.00% in MAPE, with extensive evaluations demonstrating that STeP-Diff effectively captures the spatio-temporal dependencies in air pollution fields.

Paperid: 550, https://arxiv.org/pdf/2512.03500.pdf

Abstract:
Long-form video understanding requires efficient navigation of extensive visual data to pinpoint sparse yet critical information. Current approaches to longform video understanding either suffer from severe computational overhead due to dense preprocessing, or fail to effectively balance exploration and exploitation, resulting in incomplete information coverage and inefficiency. In this work, we introduce EEA, a novel video agent framework that archives exploration-exploitation balance through semantic guidance with hierarchical tree search process. EEA autonomously discovers and dynamically updates task-relevant semantic queries, and collects video frames closely matched to these queries as semantic anchors. During the tree search process, instead of uniform expansion, EEA preferentially explores semantically relevant frames while ensuring sufficient coverage within unknown segments. Moreover, EEA adaptively combines intrinsic rewards from visionlanguage models (VLMs) with semantic priors by explicitly modeling uncertainty to achieve stable and precise evaluation of video segments. Experiments across various long-video benchmarks validate the superior performance and computational efficiency of our proposed method.

Paperid: 551, https://arxiv.org/pdf/2512.03994.pdf

Abstract:
Aligning proprietary large language models (LLMs) with internal organizational policies has become an urgent priority as organizations increasingly deploy LLMs in sensitive domains such as legal support, finance, and medical services. Beyond generic safety filters, enterprises require reliable mechanisms to detect policy violations within their regulatory and operational frameworks, where breaches can trigger legal and reputational risks. Existing content moderation frameworks, such as guardrails, remain largely confined to the safety domain and lack the robustness to capture nuanced organizational policies. LLM-as-a-judge and fine-tuning approaches, though flexible, introduce significant latency and lack interpretability. To address these limitations, we propose a training-free and efficient method that treats policy violation detection as an out-of-distribution (OOD) detection problem. Inspired by whitening techniques, we apply a linear transformation to decorrelate the model's hidden activations and standardize them to zero mean and unit variance, yielding near-identity covariance matrix. In this transformed space, we use the Euclidean norm as a compliance score to detect policy violations. The method requires only the policy text and a small number of illustrative samples, which makes it light-weight and easily deployable. On a challenging policy benchmark, our approach achieves state-of-the-art results, surpassing both existing guardrails and fine-tuned reasoning models. This work provides organizations with a practical and statistically grounded framework for policy-aware oversight of LLMs, advancing the broader goal of deployable AI governance. Code is available at: https://tinyurl.com/policy-violation-detection

Paperid: 552, https://arxiv.org/pdf/2512.00307.pdf

Abstract:
Signed graphs with positive and negative edges can model complex relationships in social networks. Leveraging on balance theory that deduces edge signs from multi-hop node pairs, signed graph learning can generate node embeddings that preserve both structural and sign information. However, training on sensitive signed graphs raises significant privacy concerns, as model parameters may leak private link information. Existing protection methods with differential privacy (DP) typically rely on edge or gradient perturbation for unsigned graph protection. Yet, they are not well-suited for signed graphs, mainly because edge perturbation tends to cascading errors in edge sign inference under balance theory, while gradient perturbation increases sensitivity due to node interdependence and gradient polarity change caused by sign flips, resulting in larger noise injection. In this paper, motivated by the robustness of adversarial learning to noisy interactions, we present ASGL, a privacy-preserving adversarial signed graph learning method that preserves high utility while achieving node-level DP. We first decompose signed graphs into positive and negative subgraphs based on edge signs, and then design a gradient-perturbed adversarial module to approximate the true signed connectivity distribution. In particular, the gradient perturbation helps mitigate cascading errors, while the subgraph separation facilitates sensitivity reduction. Further, we devise a constrained breadth-first search tree strategy that fuses with balance theory to identify the edge signs between generated node pairs. This strategy also enables gradient decoupling, thereby effectively lowering gradient sensitivity. Extensive experiments on real-world datasets show that ASGL achieves favorable privacy-utility trade-offs across multiple downstream tasks.

Paperid: 553, https://arxiv.org/pdf/2512.04864.pdf

Abstract:
Large Language Model (LLM)-based agents are increasingly used as autonomous subordinates that carry out tasks for users. This raises the question of whether they may also engage in deception, similar to how individuals in human organizations lie to superiors to create a good image or avoid punishment. We observe and define agentic upward deception, a phenomenon in which an agent facing environmental constraints conceals its failure and performs actions that were not requested without reporting. To assess its prevalence, we construct a benchmark of 200 tasks covering five task types and eight realistic scenarios in a constrained environment, such as broken tools or mismatched information sources. Evaluations of 11 popular LLMs reveal that these agents typically exhibit action-based deceptive behaviors, such as guessing results, performing unsupported simulations, substituting unavailable information sources, and fabricating local files. We further test prompt-based mitigation and find only limited reductions, suggesting that it is difficult to eliminate and highlighting the need for stronger mitigation strategies to ensure the safety of LLM-based agents.

Paperid: 554, https://arxiv.org/pdf/2512.02652.pdf

Abstract:
Existing methods for expressive music performance rendering rely on supervised learning over small labeled datasets, which limits scaling of both data volume and model size, despite the availability of vast unlabeled music, as in vision and language. To address this gap, we introduce Pianist Transformer, with four key contributions: 1) a unified Musical Instrument Digital Interface (MIDI) data representation for learning the shared principles of musical structure and expression without explicit annotation; 2) an efficient asymmetric architecture, enabling longer contexts and faster inference without sacrificing rendering quality; 3) a self-supervised pre-training pipeline with 10B tokens and 135M-parameter model, unlocking data and model scaling advantages for expressive performance rendering; 4) a state-of-the-art performance model, which achieves strong objective metrics and human-level subjective ratings. Overall, Pianist Transformer establishes a scalable path toward human-like performance synthesis in the music domain.

Paperid: 555, https://arxiv.org/pdf/2512.01848.pdf

Abstract:
Large reasoning models (LRMs) extend large language models by generating explicit chain-of-thought (CoT) reasoning, significantly improving mathematical and logical problem solving. However, this explicit reasoning process also introduces new safety risks, as unsafe behaviors often emerge within intermediate reasoning trajectories, even when final answers appear harmless. Existing safety alignment approaches primarily rely on supervised fine-tuning (SFT) over safety-oriented long CoT datasets. While intuitive, we find that SFT produces inconsistent safety improvements, degrades reasoning ability, and generalizes poorly across model families. These limitations suggest that purely supervised approaches are insufficient for robust safety alignment in LRMs. To address this, we investigate reinforcement learning (RL) as a complementary optimization framework for LRM safety training. Unlike SFT, RL directly optimizes model policies with reward feedback, enabling more adaptive and stable alignment. Extensive experiments across multiple model families and benchmarks show that RL achieves stronger and more consistent safety gains while maintaining reasoning competence. Further analysis of reflection dynamics and token-level entropy reveals that RL suppresses unsafe exploratory reasoning while preserving reflective depth, leading to safer and more reliable reasoning processes.

Paperid: 556, https://arxiv.org/pdf/2512.04550.pdf

Abstract:
The quadratic complexity of self-attention constrains Large Language Models (LLMs) in processing long contexts, a capability essential for many advanced applications. Context compression aims to alleviate this computational bottleneck while retaining critical semantic information. However, existing approaches often fall short: explicit methods may compromise local detail, whereas implicit methods can suffer from positional biases, information degradation, or an inability to capture long-range semantic dependencies. We propose AdmTree, a novel framework for adaptive, hierarchical context compression with a central focus on preserving high semantic fidelity while maintaining efficiency. AdmTree dynamically segments input based on information density, utilizing gist tokens to summarize variable-length segments as the leaves of a semantic binary tree. This structure, together with a lightweight aggregation mechanism and a frozen backbone LLM (thereby minimizing new trainable parameters), enables efficient hierarchical abstraction of the context. By preserving fine-grained details alongside global semantic coherence, mitigating positional bias, and dynamically adapting to content, AdmTree robustly retains the semantic information of long contexts.

Paperid: 557, https://arxiv.org/pdf/2512.07596.pdf

Abstract:
The recent Segment Anything Model (SAM) 3 has introduced significant advancements over its predecessor, SAM 2, particularly with the integration of language-based segmentation and enhanced 3D perception capabilities. SAM 3 supports zero-shot segmentation across a wide range of prompts, including point, bounding box, and language-based prompts, allowing for more flexible and intuitive interactions with the model. In this empirical evaluation, we assess the performance of SAM 3 in robot-assisted surgery, benchmarking its zero-shot segmentation with point and bounding box prompts and exploring its effectiveness in dynamic video tracking, alongside its newly introduced language prompt segmentation. While language prompts show potential, their performance in the surgical domain is currently suboptimal, highlighting the need for further domain-specific training. Additionally, we investigate SAM 3's 3D reconstruction abilities, demonstrating its capacity to process surgical scene data and reconstruct 3D anatomical structures from 2D images. Through comprehensive testing on the MICCAI EndoVis 2017 and EndoVis 2018 benchmarks, SAM 3 shows clear improvements over SAM and SAM 2 in both image and video segmentation under spatial prompts, while zero-shot evaluations on SCARED, StereoMIS, and EndoNeRF indicate strong monocular depth estimation and realistic 3D instrument reconstruction, yet also reveal remaining limitations in complex, highly dynamic surgical scenes.

Paperid: 558, https://arxiv.org/pdf/2512.06921.pdf

Abstract:
Multimodal Large Language Models (MLLMs) have shown significant potential in surgical video understanding. With improved zero-shot performance and more effective human-machine interaction, they provide a strong foundation for advancing surgical education and assistance. However, existing research and datasets primarily focus on understanding surgical procedures and workflows, while paying limited attention to the critical role of anatomical comprehension. In clinical practice, surgeons rely heavily on precise anatomical understanding to interpret, review, and learn from surgical videos. To fill this gap, we introduce the Neurosurgical Anatomy Benchmark (NeuroABench), the first multimodal benchmark explicitly created to evaluate anatomical understanding in the neurosurgical domain. NeuroABench consists of 9 hours of annotated neurosurgical videos covering 89 distinct procedures and is developed using a novel multimodal annotation pipeline with multiple review cycles. The benchmark evaluates the identification of 68 clinical anatomical structures, providing a rigorous and standardized framework for assessing model performance. Experiments on over 10 state-of-the-art MLLMs reveal significant limitations, with the best-performing model achieving only 40.87% accuracy in anatomical identification tasks. To further evaluate the benchmark, we extract a subset of the dataset and conduct an informative test with four neurosurgical trainees. The results show that the best-performing student achieves 56% accuracy, with the lowest scores of 28% and an average score of 46.5%. While the best MLLM performs comparably to the lowest-scoring student, it still lags significantly behind the group's average performance. This comparison underscores both the progress of MLLMs in anatomical understanding and the substantial gap that remains in achieving human-level performance.

Paperid: 559, https://arxiv.org/pdf/2512.06859.pdf

Abstract:
In this work, we present JT-DA-8B (JiuTian Data Analyst 8B), a specialized large language model designed for complex table reasoning tasks across diverse real-world scenarios. To address the lack of high-quality supervision in tabular reasoning scenarios, we construct a comprehensive and diverse training corpus with 34 well-defined table reasoning tasks, by aggregating 29 public table QA datasets and 3 million tables. An automatic pipeline is proposed to generate realistic multi-step analytical tasks involving reasoning patterns. The model is trained upon open-source JT-Coder-8B model, an 8B-parameter decoder-only foundation model trained from scratch. In the training stage, we leverage LLM-based scoring and workflow-aligned filtering to distill high-quality, table-centric data. Both supervised fine-tuning (SFT) and Reinforcement learning (RL) are adopted to optimize our model. Afterwards, a four-stage table reasoning workflow is proposed, including table preprocessing, table sensing, tool-integrated reasoning, and prompt engineering, to improve model interpretability and execution accuracy. Experimental results show that JT-DA-8B achieves strong performance in various table reasoning tasks, demonstrating the effectiveness of data-centric generation and workflow-driven optimization.

Paperid: 560, https://arxiv.org/pdf/2512.03627.pdf

Abstract:
Despite rapid progress in large-scale language and vision models, AI agents still suffer from a fundamental limitation: they cannot remember. Without reliable memory, agents catastrophically forget past experiences, struggle with long-horizon reasoning, and fail to operate coherently in multimodal or interactive environments. We introduce MemVerse, a model-agnostic, plug-and-play memory framework that bridges fast parametric recall with hierarchical retrieval-based memory, enabling scalable and adaptive multimodal intelligence. MemVerse maintains short-term memory for recent context while transforming raw multimodal experiences into structured long-term memories organized as hierarchical knowledge graphs. This design supports continual consolidation, adaptive forgetting, and bounded memory growth. To handle real-time demands, MemVerse introduces a periodic distillation mechanism that compresses essential knowledge from long-term memory into the parametric model, allowing fast, differentiable recall while preserving interpretability. Extensive experiments demonstrate that MemVerse significantly improves multimodal reasoning and continual learning efficiency, empowering agents to remember, adapt, and reason coherently across extended interactions.

Paperid: 561, https://arxiv.org/pdf/2512.03542.pdf

Abstract:
Multimodal Large Language Models (MLLMs) excel in numerous vision-language tasks yet suffer from hallucinations, producing content inconsistent with input visuals, that undermine reliability in precision-sensitive domains. This issue stems from a fundamental problem of visual neglect, where models fail to adequately prioritize input images. Existing methods typically alleviate hallucinations by intervening in the attention score or output logits, focusing on "how to intervene" but overlooking the prerequisite "when to intervene", which leads to the "over-intervention" problem and subsequently introduces new hallucinations and unnecessary computational overhead. To address this gap, we first investigate the mechanism of visual neglect and reveal it can be accurately detected via head-level activation patterns in MLLMs. We thus propose V-ITI, a lightweight visual inference-time intervention framework integrating a Visual Neglect Detector that identifies visual neglect via head-level discriminative probes and a Visual Recall Intervenor that modulates activations with prestored visual activation information only when the visual neglect is detected. Extensive experiments across eight benchmarks and different MLLM families demonstrate that V-ITI consistently mitigates vision-related hallucinations while preserving general task performance.

Paperid: 562, https://arxiv.org/pdf/2512.03405.pdf

Abstract:
Understanding visual differences between dynamic scenes requires the comparative perception of compositional, spatial, and temporal changes--a capability that remains underexplored in existing vision-language systems. While prior work on Image Difference Captioning (IDC) has enabled models to describe semantic changes between static images, these approaches fail to capture motion continuity, event evolution, or editing consistency over time. We introduce the ViDiC (Video Difference Captioning) task and its corresponding ViDiC-1K dataset, designed to evaluate the ability of Multimodal Large Language Models (MLLMs) to provide fine-grained descriptions of similarities and differences between video pairs. ViDiC-1K comprises 1,000 curated video pairs annotated with over 4,000 comparative checklist items, covering seven categories: subject, style, background, cinematography, motion, location, and playback techniques. To ensure reliable evaluation, we propose a dual-checklist framework that measures the accuracy of similarity and difference separately, based on the LLM-as-a-Judge protocol. Experiments on nineteen representative multimodal models reveal a significant performance gap in their comparative description and difference perception abilities. We hope ViDiC-1K can be a challenging benchmark that lays a solid foundation for advancing video understanding, edit awareness, and comparative reasoning in multimodal intelligence.

Paperid: 563, https://arxiv.org/pdf/2512.00255.pdf

Abstract:
We present Relightable Holoported Characters (RHC), a novel person-specific method for free-view rendering and relighting of full-body and highly dynamic humans solely observed from sparse-view RGB videos at inference. In contrast to classical one-light-at-a-time (OLAT)-based human relighting, our transformer-based RelightNet predicts relit appearance within a single network pass, avoiding costly OLAT-basis capture and generation. For training such a model, we introduce a new capture strategy and dataset recorded in a multi-view lightstage, where we alternate frames lit by random environment maps with uniformly lit tracking frames, simultaneously enabling accurate motion tracking and diverse illumination as well as dynamics coverage. Inspired by the rendering equation, we derive physics-informed features that encode geometry, albedo, shading, and the virtual camera view from a coarse human mesh proxy and the input views. Our RelightNet then takes these features as input and cross-attends them with a novel lighting condition, and regresses the relit appearance in the form of texel-aligned 3D Gaussian splats attached to the coarse mesh proxy. Consequently, our RelightNet implicitly learns to efficiently compute the rendering equation for novel lighting conditions within a single feed-forward pass. Experiments demonstrate our method's superior visual fidelity and lighting reproduction compared to state-of-the-art approaches. Project page: https://vcai.mpi-inf.mpg.de/projects/RHC/

Paperid: 564, https://arxiv.org/pdf/2512.07189.pdf

Abstract:
Decentralized Storage Networks (DSNs) are emerging as a foundational infrastructure for Web 3.0, offering global peer-to-peer storage. However, a critical vulnerability persists: user privacy during file retrieval remains largely unaddressed, risking the exposure of sensitive information. To overcome this, we introduce PIR-DSN, the first DSN protocol to integrate Private Information Retrieval (PIR) for both single and multi-server settings. Our key innovations include a novel secure mapping method that transforms sparse file identifiers into compact integer indexes, enabling both public verifiability of file operations and efficient private retrieval. Furthermore, PIR-DSN guarantees Byzantine-robust private retrieval through file replication across multiple miners. We implement and rigorously evaluate PIR-DSN against three prominent industrial DSN systems. Experimental results demonstrate that PIR-DSN achieves comparable overhead for file upload and deletion. While PIR inherently introduces an additional computational cost leading to higher retrieval latency, PIR-DSN maintains comparable throughput. These findings underscore PIR-DSN's practical viability for privacy-sensitive applications within DSN environments.

Paperid: 565, https://arxiv.org/pdf/2512.06387.pdf

Abstract:
Embodied AI systems integrate language models with real world sensing, mobility, and cloud connected mobile apps. Yet while model jailbreaks have drawn significant attention, the broader system stack of embodied intelligence remains largely unexplored. In this work, we conduct the first holistic security analysis of the Unitree Go2 platform and uncover ten cross layer vulnerabilities the "Ten Sins of Embodied AI Security." Using BLE sniffing, traffic interception, APK reverse engineering, cloud API testing, and hardware probing, we identify systemic weaknesses across three architectural layers: wireless provisioning, core modules, and external interfaces. These include hard coded keys, predictable handshake tokens, WiFi credential leakage, missing TLS validation, static SSH password, multilingual safety bypass behavior, insecure local relay channels, weak binding logic, and unrestricted firmware access. Together, they allow adversaries to hijack devices, inject arbitrary commands, extract sensitive information, or gain full physical control.Our findings show that securing embodied AI requires far more than aligning the model itself. We conclude with system level lessons learned and recommendations for building embodied platforms that remain robust across their entire software hardware ecosystem.

Paperid: 566, https://arxiv.org/pdf/2512.03791.pdf

Abstract:
Cross-chain technology enables interoperability among otherwise isolated blockchains, supporting interactions across heterogeneous networks. Similar to how multi-hop communication became fundamental in the evolution of the Internet, the demand for multi-hop cross-chain interactions is gaining increasing attention. However, this growing demand introduces new security and privacy challenges. On the security side, multi-hop interactions depend on the availability of multiple participating nodes. If any node becomes temporarily offline during execution, the protocol may fail to complete correctly, leading to settlement failure or fund loss. On the privacy side, the need for on-chain transparency to validate intermediate states may unintentionally leak linkable information, compromising the unlinkability of user interactions. In this paper, we propose the Cross-Chain Channel Network (CCN), a decentralized network designed to support secure and privacy-preserving multi-hop cross-chain transactions. Through experimental evaluation, we identify two critical types of offline failures, referred to as active and passive offline cases, which have not been adequately addressed by existing solutions. To mitigate these issues, we introduce R-HTLC, a core protocol within CCN. R-HTLC incorporates an hourglass mechanism and a multi-path refund strategy to ensure settlement correctness even when some nodes go offline during execution. Importantly, CCN addresses not only the correctness under offline conditions but also maintains unlinkability in such adversarial settings. To overcome this, CCN leverages zero-knowledge proofs and off-chain coordination, ensuring that interaction relationships remain indistinguishable even when certain nodes are temporarily offline.

Paperid: 567, https://arxiv.org/pdf/2512.03775.pdf

Abstract:
The Model Context Protocol (MCP) is rapidly emerging as the middleware for LLM-based applications, offering a standardized interface for tool integration. However, its built-in security mechanisms are minimal: while schemas and declarations prevent malformed requests, MCP provides no guarantees of authenticity or confidentiality, forcing developers to implement cryptography themselves. Such ad hoc practices are historically prone to misuse, and within MCP they threaten sensitive data and services. We present MICRYSCOPE, the first domain-specific framework for detecting cryptographic misuses in MCP implementations. MICRYSCOPE combines three key innovations: a cross-language intermediate representation that normalizes cryptographic APIs across diverse ecosystems, a hybrid dependency analysis that uncovers explicit and implicit function relationships (including insecure runtime compositions orchestrated by LLMs) and a taint-based misuse detector that tracks sensitive data flows and flags violations of established cryptographic rules. Applying MICRYSCOPE to 9,403 MCP servers, we identified 720 with cryptographic logic, of which 19.7% exhibited misuses. These flaws are concentrated in certain markets (e.g., Smithery Registry with 42% insecure servers), languages (Python at 34% misuse rate), and categories (Developer Tools and Data Science & ML accounting for over 50% of all misuses). Case studies reveal real-world consequences, including leaked API keys, insecure DES/ECB tools, and MD5-based authentication bypasses. Our study establishes the first ecosystem-wide view of cryptographic misuse in MCP and provides both tools and insights to strengthen the security foundations of this rapidly growing protocol.

Paperid: 568, https://arxiv.org/pdf/2512.01485.pdf

Abstract:
Chain-of-Thought (CoT) reasoning has significantly advanced the problem-solving capabilities of Large Language Models (LLMs), yet conventional CoT often exhibits internal determinism during decoding, limiting exploration of plausible alternatives. Recent methods attempt to address this by generating soft abstract tokens to enable reasoning in a continuous semantic space. However, we find that such approaches remain constrained by the greedy nature of autoregressive decoding, which fundamentally isolates the model from alternative reasoning possibilities. In this work, we propose Multi-Path Perception Policy Optimization (M3PO), a novel reinforcement learning framework that explicitly injects collective insights into the reasoning process. M3PO leverages parallel policy rollouts as naturally diverse reasoning sources and integrates cross-path interactions into policy updates through a lightweight collaborative mechanism. This design allows each trajectory to refine its reasoning with peer feedback, thereby cultivating more reliable multi-step reasoning patterns. Empirical results show that M3PO achieves state-of-the-art performance on both knowledge- and reasoning-intensive benchmarks. Models trained with M3PO maintain interpretability and inference efficiency, underscoring the promise of multi-path collaborative learning for robust reasoning.

Paperid: 569, https://arxiv.org/pdf/2512.01442.pdf

Abstract:
Multimodal sentiment analysis (MSA) is a research field that recognizes human sentiments by combining textual, visual, and audio modalities. The main challenge lies in integrating sentiment-related information from different modalities, which typically arises during the unimodal feature extraction phase and the multimodal feature fusion phase. Existing methods extract only shallow information from unimodal features during the extraction phase, neglecting sentimental differences across different personalities. During the fusion phase, they directly merge the feature information from each modality without considering differences at the feature level. This ultimately affects the model's recognition performance. To address this problem, we propose a personality-sentiment aligned multi-level fusion framework. We introduce personality traits during the feature extraction phase and propose a novel personality-sentiment alignment method to obtain personalized sentiment embeddings from the textual modality for the first time. In the fusion phase, we introduce a novel multi-level fusion method. This method gradually integrates sentimental information from textual, visual, and audio modalities through multimodal pre-fusion and a multi-level enhanced fusion strategy. Our method has been evaluated through multiple experiments on two commonly used datasets, achieving state-of-the-art results.

Paperid: 570, https://arxiv.org/pdf/2512.00903.pdf

Abstract:
Vision-Language-Action (VLA) models built on pretrained Vision-Language Models (VLMs) show strong potential but are limited in practicality due to their large parameter counts. To mitigate this issue, using a lightweight VLM has been explored, but it compromises spatiotemporal reasoning. Although some methods suggest that incorporating additional 3D inputs can help, they usually rely on large VLMs to fuse 3D and 2D inputs and still lack temporal understanding. Therefore, we propose SwiftVLA, an architecture that enhances a compact model with 4D understanding while preserving design efficiency. Specifically, our approach features a pretrained 4D visual geometry transformer with a temporal cache that extracts 4D features from 2D images. Then, to enhance the VLM's ability to exploit both 2D images and 4D features, we introduce Fusion Tokens, a set of learnable tokens trained with a future prediction objective to generate unified representations for action generation. Finally, we introduce a mask-and-reconstruct strategy that masks 4D inputs to the VLM and trains the VLA to reconstruct them, enabling the VLM to learn effective 4D representations and allowing the 4D branch to be dropped at inference with minimal performance loss. Experiments in real and simulated environments show that SwiftVLA outperforms lightweight baselines and rivals VLAs up to 7 times larger, achieving comparable performance on edge devices while being 18 times faster and reducing memory footprint by 12 times.

Paperid: 571, https://arxiv.org/pdf/2512.06865.pdf

Abstract:
Existing autonomous driving systems rely on onboard sensors (cameras, LiDAR, IMU, etc) for environmental perception. However, this paradigm is limited by the drive-time perception horizon and often fails under limited view scope, occlusion or extreme conditions such as darkness and rain. In contrast, human drivers are able to recall road structure even under poor visibility. To endow models with this ``recall" ability, we propose the spatial retrieval paradigm, introducing offline retrieved geographic images as an additional input. These images are easy to obtain from offline caches (e.g, Google Maps or stored autonomous driving datasets) without requiring additional sensors, making it a plug-and-play extension for existing AD tasks. For experiments, we first extend the nuScenes dataset with geographic images retrieved via Google Maps APIs and align the new data with ego-vehicle trajectories. We establish baselines across five core autonomous driving tasks: object detection, online mapping, occupancy prediction, end-to-end planning, and generative world modeling. Extensive experiments show that the extended modality could enhance the performance of certain tasks. We will open-source dataset curation code, data, and benchmarks for further study of this new autonomous driving paradigm.

Paperid: 572, https://arxiv.org/pdf/2512.06275.pdf

Abstract:
Vital sign measurement using cameras presents opportunities for comfortable, ubiquitous health monitoring. Remote photoplethysmography (rPPG), a foundational technology, enables cardiac measurement through minute changes in light reflected from the skin. However, practical deployment is limited by the computational constraints of performing analysis on front-end devices and the accuracy degradation of transmitting data through compressive channels that reduce signal quality. We propose a memory efficient rPPG algorithm - \emph{FacePhys} - built on temporal-spatial state space duality, which resolves the trilemma of model scalability, cross-dataset generalization, and real-time operation. Leveraging a transferable heart state, FacePhys captures subtle periodic variations across video frames while maintaining a minimal computational overhead, enabling training on extended video sequences and supporting low-latency inference. FacePhys establishes a new state-of-the-art, with a substantial 49\% reduction in error. Our solution enables real-time inference with a memory footprint of 3.6 MB and per-frame latency of 9.46 ms -- surpassing existing methods by 83\% to 99\%. These results translate into reliable real-time performance in practical deployments, and a live demo is available at https://www.facephys.com/.

Paperid: 573, https://arxiv.org/pdf/2512.02657.pdf

Abstract:
The recent rapid growth of visual generative models trained on vast web-scale datasets has created significant tension with data privacy regulations and copyright laws, such as GDPR's ``Right to be Forgotten.'' This necessitates machine unlearning (MU) to remove specific concepts without the prohibitive cost of retraining. However, existing MU techniques are fundamentally ill-equipped for real-world scenarios where deletion requests arrive sequentially, a setting known as continual unlearning (CUL). Naively applying one-shot methods in a continual setting triggers a stability crisis, leading to a cascade of degradation characterized by retention collapse, compounding collateral damage to related concepts, and a sharp decline in generative quality. To address this critical challenge, we introduce a novel generative distillation based continual unlearning framework that ensures targeted and stable unlearning under sequences of deletion requests. By reframing each unlearning step as a multi-objective, teacher-student distillation process, the framework leverages principles from continual learning to maintain model integrity. Experiments on a 10-step sequential benchmark demonstrate that our method unlearns forget concepts with better fidelity and achieves this without significant interference to the performance on retain concepts or the overall image quality, substantially outperforming baselines. This framework provides a viable pathway for the responsible deployment and maintenance of large-scale generative models, enabling industries to comply with ongoing data removal requests in a practical and effective manner.

Paperid: 574, https://arxiv.org/pdf/2512.00756.pdf

Abstract:
With the advancement of computational resources, Large Vision-Language Models (LVLMs) exhibit impressive Perception and Reasoning (P&R) performance on Graphical User Interface (GUI) tasks. However, although they demonstrate strong P&R capabilities in English GUI scenarios, their performance in multilingual settings has received little attention, which limits their global applications. Moreover, existing studies on GUI tasks lack fine-grained analyses, including widget functions and elements' spatial relationships, which are fundamental for more targeted improvements. To tackle these issues, we propose MPR-GUI-Bench, a Multilingual fine-grained Perception and Reasoning GUI Benchmark to evaluate GUI agents' P&R capabilities. Evaluation results demonstrate that LVLMs exhibit significantly worse P&R performance in non-English languages than in English. To address these gaps, we propose GUI-XLI, a GUI Cross-Lingual Intervention method that applies interventions to the hidden states at P&R capability-related layers to mitigate the gaps between English and other languages, building on previous research showing that the hidden states of different language inputs exhibit significant differences in the latent space. Experimental results indicate that our method improves GUI agents' multilingual P&R capability by 6.5% on average.

Paperid: 575, https://arxiv.org/pdf/2512.07175.pdf

Abstract:
Self-play fine-tuning has demonstrated promising abilities in adapting large language models (LLMs) to downstream tasks with limited real-world data. The basic principle is to iteratively refine the model with real samples and synthetic ones generated from itself. However, the existing methods primarily focus on the relative gaps between the rewards for two types of data, neglecting their absolute values. Through theoretical analysis, we identify that the gap-based methods suffer from unstable evolution, due to the potentially degenerated objectives. To address this limitation, we introduce a novel self-play fine-tuning method, namely Self-PlAy via Noise Contrastive Estimation (SPACE), which leverages noise contrastive estimation to capture the real-world data distribution. Specifically, SPACE treats synthetic samples as auxiliary components, and discriminates them from the real ones in a binary classification manner. As a result, SPACE independently optimizes the absolute reward values for each type of data, ensuring a consistently meaningful objective and thereby avoiding the instability issue. Theoretically, we show that the optimal solution of the objective in SPACE aligns with the underlying distribution of real-world data, and SPACE guarantees a provably stable convergence to the optimal distribution. Empirically, we show that SPACE significantly improves the performance of LLMs over various tasks, and outperforms supervised fine-tuning that employs much more real-world samples. Compared to gap-based self-play fine-tuning methods, SPACE exhibits remarkable superiority and stable evolution.

Paperid: 576, https://arxiv.org/pdf/2512.06653.pdf

Abstract:
DeepSearch paradigms have become a core enabler for deep reasoning models, allowing them to invoke external search tools to access up-to-date, domain-specific knowledge beyond parametric boundaries, thereby enhancing the depth and factual reliability of reasoning. Building upon this foundation, recent advances in reinforcement learning (RL) have further empowered models to autonomously and strategically control search tool usage, optimizing when and how to query external knowledge sources. Yet, these RL-driven DeepSearch systems often reveal a see-saw trade-off between accuracy and efficiency-frequent tool invocations can improve factual correctness but lead to unnecessary computational overhead and diminished efficiency. To address this challenge, we propose LightSearcher, an efficient RL framework that incorporates textual experiential memory by learning contrastive reasoning trajectories to generate interpretable summaries of successful reasoning patterns. In addition, it employs an adaptive reward shaping mechanism that penalizes redundant tool calls only in correct-answer scenarios. This design effectively balances the inherent accuracy-efficiency trade-off in DeepSearch paradigms. Experiments on four multi-hop QA benchmarks show that LightSearcher maintains accuracy comparable to SOTA baseline ReSearch, while reducing search tool invocations by 39.6%, inference time by 48.6%, and token consumption by 21.2%, demonstrating its superior efficiency.

Paperid: 577, https://arxiv.org/pdf/2512.01316.pdf

Abstract:
Minimum Bayes risk (MBR) decoding generates high-quality translations by maximizing the expected utility of output candidates, but it evaluates all pairwise scores over the candidate set; hence, it takes quadratic time with respect to the number of candidates. To reduce the number of utility function calls, probabilistic MBR (PMBR) decoding partially evaluates quality scores using sampled pairs of candidates and completes the missing scores with a matrix completion algorithm. Nevertheless, it degrades the translation quality as the number of utility function calls is reduced. Therefore, to improve the trade-off between quality and cost, we propose agreement-constrained PMBR (AC-PMBR) decoding, which leverages a knowledge distilled model to guide the completion of the score matrix. Our AC-PMBR decoding improved approximation errors of matrix completion by up to 3 times and achieved higher translation quality compared with PMBR decoding at a comparable computational cost on the WMT'23 En$\leftrightarrow$De translation tasks.

Paperid: 578, https://arxiv.org/pdf/2512.00778.pdf

Abstract:
Preference optimization (PO) is indispensable for large language models (LLMs), with methods such as direct preference optimization (DPO) and proximal policy optimization (PPO) achieving great success. A common belief is that DPO is supervised learning while PPO is reinforcement learning, yet deeper analyses for the reasons underlying these differences remain lacking. To fill this gap, we analyze their optimization dynamics, revealing distinct algorithmic behaviors and comprehending their underlying causes. First, we examine the target directions of gradient-based updates and find that DPO follows stable targets, whereas PPO follows dynamic targets that balance exploration and exploitation, thus validating the common belief from a new perspective. Second, we examine the roles of positive learning, negative learning, and loss reweighting, which are three key components in PO methods. Our analyses reveal that these components play fairly different roles. In DPO, positive and negative learning jointly shape the learning targets meanwhile mutually offset each other. However, loss reweighting in DPO acts less as a reward signal but more as a regularizer to mitigate overfitting. In PPO, negative learning primarily supports exploration rather than determining the targets. Meanwhile, loss reweighting, related to absolute values of token-level advantages, indicates the distinct roles of token groups in updating targets. Given these findings, we conduct carefully designed ablation studies to further examine how controlling these dynamics impacts optimization efficiency and practical performance. The insights gained from our analyses not only deepen the understanding of PO methods but also inspire the development of more preference-aligned LLMs.

Paperid: 579, https://arxiv.org/pdf/2512.00729.pdf

Abstract:
Large reasoning models (LRMs) have garnered significant attention from researchers owing to their exceptional capability in addressing complex tasks. Motivated by the observed human-like behaviors in their reasoning processes, this paper introduces a comprehensive taxonomy to characterize atomic reasoning steps and probe the ``psyche'' of LRM intelligence. Specifically, it comprises five groups and seventeen categories derived from human mental processes, thereby grounding the understanding of LRMs in an interdisciplinary perspective. The taxonomy is then applied for an in-depth understanding of current LRMs, resulting in a distinct labeled dataset that comprises 277,534 atomic reasoning steps. Using this resource, we analyze contemporary LRMs and distill several actionable takeaways for improving training and post-training of reasoning models. Notably, our analysis reveals that prevailing post-answer ``double-checks'' (self-monitoring evaluations) are largely superficial and rarely yield substantive revisions. Thus, incentivizing comprehensive multi-step reflection, rather than simple self-monitoring, may offer a more effective path forward. To complement the taxonomy, an automatic annotation framework, named CAPO, is proposed to leverage large language models (LLMs) for generating the taxonomy-based annotations. Experimental results demonstrate that CAPO achieves higher consistency with human experts compared to baselines, facilitating a scalable and comprehensive analysis of LRMs from a human cognitive perspective. Together, the taxonomy, CAPO, and the derived insights provide a principled, scalable path toward understanding and advancing LRM reasoning.

Paperid: 580, https://arxiv.org/pdf/2512.00391.pdf

Abstract:
Model merging has emerged as a practical paradigm for integrating multiple independently trained models into a single model without joint retraining. Previous studies have demonstrated the effectiveness of combining parameters through strategies such as parameter decomposition, coefficient optimization, and subspace learning, significantly reducing the need for expensive joint training and achieving strong empirical performance across diverse tasks. However, these approaches predominantly treat merging as a problem of parameter space decomposition or fusion coefficient optimization, while overlooking the critical role of directional information in both parameter and feature spaces. In practice, naïve merging introduces inconsistencies in dominant parameter directions and disrupts structural coherence across models, which can degrade performance. Moreover, coefficient-based optimization methods implicitly assume compatible feature-space directions across models. However, Neural Collapse indicates that class features follow structured directional patterns, which may differ across independently trained models, making coefficient optimization alone insufficient. In this work, we emphasize the importance of \emph{directional alignment} and introduce a unified geometric framework, \emph{Merging with Directional Alignment} (\method{}), which aligns directional structures consistently in both the parameter and feature spaces. Our analysis shows that directional alignment improves structural coherence, and extensive experiments across benchmarks, model scales, and task configurations further validate the effectiveness of our approach.

Paperid: 581, https://arxiv.org/pdf/2512.00349.pdf

Abstract:
Are frontier AI systems becoming more capable? Certainly. Yet such progress is not an unalloyed blessing but rather a Trojan horse: behind their performance leaps lie more insidious and destructive safety risks, namely deception. Unlike hallucination, which arises from insufficient capability and leads to mistakes, deception represents a deeper threat in which models deliberately mislead users through complex reasoning and insincere responses. As system capabilities advance, deceptive behaviours have spread from textual to multimodal settings, amplifying their potential harm. First and foremost, how can we monitor these covert multimodal deceptive behaviors? Nevertheless, current research remains almost entirely confined to text, leaving the deceptive risks of multimodal large language models unexplored. In this work, we systematically reveal and quantify multimodal deception risks, introducing MM-DeceptionBench, the first benchmark explicitly designed to evaluate multimodal deception. Covering six categories of deception, MM-DeceptionBench characterizes how models strategically manipulate and mislead through combined visual and textual modalities. On the other hand, multimodal deception evaluation is almost a blind spot in existing methods. Its stealth, compounded by visual-semantic ambiguity and the complexity of cross-modal reasoning, renders action monitoring and chain-of-thought monitoring largely ineffective. To tackle this challenge, we propose debate with images, a novel multi-agent debate monitor framework. By compelling models to ground their claims in visual evidence, this method substantially improves the detectability of deceptive strategies. Experiments show that it consistently increases agreement with human judgements across all tested models, boosting Cohen's kappa by 1.5x and accuracy by 1.25x on GPT-4o.

Paperid: 582, https://arxiv.org/pdf/2512.00305.pdf

Abstract:
Multimodal Large Language Models (MLLMs) have emerged as powerful tools for chart comprehension. However, they heavily rely on extracted content via OCR, which leads to numerical hallucinations when chart textual annotations are sparse. While existing methods focus on scaling instructions, they fail to address the fundamental challenge, i.e., reasoning with visual perception. In this paper, we identify a critical observation: MLLMs exhibit weak grounding in chart elements and proportional relationships, as evidenced by their inability to localize key positions to match their reasoning. To bridge this gap, we propose PointCoT, which integrates reflective interaction into chain-of-thought reasoning in charts. By prompting MLLMs to generate bounding boxes and re-render charts based on location annotations, we establish connections between textual reasoning steps and visual grounding regions. We further introduce an automated pipeline to construct ChartPoint-SFT-62k, a dataset featuring 19.2K high-quality chart samples with step-by-step CoT, bounding box, and re-rendered visualizations. Leveraging this data, we develop two instruction-tuned models, ChartPointQ2 and ChartPointQ2.5, which outperform state-of-the-art across several chart benchmarks, e.g., +5.04\% on ChartBench.

Paperid: 583, https://arxiv.org/pdf/2512.06835.pdf

Abstract:
Recent vision-language models (VLMs) achieve remarkable reasoning through reinforcement learning (RL), which provides a feasible solution for realizing continuous self-evolving large vision-language models (LVLMs) in the era of experience. However, RL for VLMs requires abundant high-quality multimodal data, especially challenging in specialized domains like chemistry, earth sciences, and multimodal mathematics. Existing strategies such as synthetic data and self-rewarding mechanisms suffer from limited distributions and alignment difficulties, ultimately causing reward hacking: models exploit high-reward patterns, collapsing policy entropy and destabilizing training. We propose DoGe (Decouple to Generalize), a dual-decoupling framework that guides models to first learn from context rather than problem solving by refocusing on the problem context scenarios overlooked by synthetic data methods. By decoupling learning process into dual components (Thinker and Solver), we reasonably quantify the reward signals of this process and propose a two-stage RL post-training approach from freely exploring context to practically solving tasks. Second, to increase the diversity of training data, DoGe constructs an evolving curriculum learning pipeline: an expanded native domain knowledge corpus and an iteratively evolving seed problems pool. Experiments show that our method consistently outperforms the baseline across various benchmarks, providing a scalable pathway for realizing self-evolving LVLMs.

Paperid: 584, https://arxiv.org/pdf/2512.04759.pdf

Authors:Malvina Nissim, Danilo Croce, Viviana Patti, Pierpaolo Basile, Giuseppe Attanasio, Elio Musacchio, Matteo Rinaldi, Federico Borazio, Maria Francis, Jacopo Gili, Daniel Scalena, Begoña Altuna, Ekhi Azurmendi, Valerio Basile, Luisa Bentivogli, Arianna Bisazza, Marianna Bolognesi, Dominique Brunato, Tommaso Caselli, Silvia Casola, Maria Cassese, Mauro Cettolo, Claudia Collacciani, Leonardo De Cosmo, Maria Pia Di Buono, Andrea Esuli, Julen Etxaniz, Chiara Ferrando, Alessia Fidelangeli, Simona Frenda, Achille Fusco, Marco Gaido, Andrea Galassi, Federico Galli, Luca Giordano, Mattia Goffetti, Itziar Gonzalez-Dios, Lorenzo Gregori, Giulia Grundler, Sandro Iannaccone, Chunyang Jiang, Moreno La Quatra, Francesca Lagioia, Soda Marem Lo, Marco Madeddu, Bernardo Magnini, Raffaele Manna, Fabio Mercorio, Paola Merlo, Arianna Muti, Vivi Nastase, Matteo Negri, Dario Onorati, Elena Palmieri, Sara Papi, Lucia Passaro, Giulia Pensa, Andrea Piergentili, Daniele Potertì, Giovanni Puccetti, Federico Ranaldi, Leonardo Ranaldi, Andrea Amelio Ravelli, Martina Rosola, Elena Sofia Ruzzetti, Giuseppe Samo, Andrea Santilli, Piera Santin, Gabriele Sarti, Giovanni Sartor, Beatrice Savoldi, Antonio Serino, Andrea Seveso, Lucia Siciliani, Paolo Torroni, Rossella Varvara, Andrea Zaninello, Asya Zanollo, Fabio Massimo Zanzotto, Kamyar Zeinalipour, Andrea Zugarini

Abstract:
The rapid progress of Large Language Models (LLMs) has transformed natural language processing and broadened its impact across research and society. Yet, systematic evaluation of these models, especially for languages beyond English, remains limited. "Challenging the Abilities of LAnguage Models in ITAlian" (CALAMITA) is a large-scale collaborative benchmarking initiative for Italian, coordinated under the Italian Association for Computational Linguistics. Unlike existing efforts that focus on leaderboards, CALAMITA foregrounds methodology: it federates more than 80 contributors from academia, industry, and the public sector to design, document, and evaluate a diverse collection of tasks, covering linguistic competence, commonsense reasoning, factual consistency, fairness, summarization, translation, and code generation. Through this process, we not only assembled a benchmark of over 20 tasks and almost 100 subtasks, but also established a centralized evaluation pipeline that supports heterogeneous datasets and metrics. We report results for four open-weight LLMs, highlighting systematic strengths and weaknesses across abilities, as well as challenges in task-specific evaluation. Beyond quantitative results, CALAMITA exposes methodological lessons: the necessity of fine-grained, task-representative metrics, the importance of harmonized pipelines, and the benefits and limitations of broad community engagement. CALAMITA is conceived as a rolling benchmark, enabling continuous integration of new tasks and models. This makes it both a resource -- the most comprehensive and diverse benchmark for Italian to date -- and a framework for sustainable, community-driven evaluation. We argue that this combination offers a blueprint for other languages and communities seeking inclusive and rigorous LLM evaluation practices.

Paperid: 585, https://arxiv.org/pdf/2512.03671.pdf

Abstract:
The rise of Artificial Intelligence (AI) language technologies, particularly generative AI (GenAI) chatbots accessible via conversational interfaces, is transforming digital interactions. While these tools hold societal promise, they also risk widening digital divides due to uneven adoption and low awareness of their limitations. This study presents the first comprehensive empirical mapping of GenAI adoption, usage patterns, and literacy in Italy, based on newly collected survey data from 1,906 Italian-speaking adults. Our findings reveal widespread adoption for both work and personal use, including sensitive tasks like emotional support and medical advice. Crucially, GenAI is supplanting other technologies to become a primary information source: this trend persists despite low user digital literacy, posing a risk as users struggle to recognize errors or misinformation. Moreover, we identify a significant gender divide -- particularly pronounced in older generations -- where women are half as likely to adopt GenAI and use it less frequently than men. While we find literacy to be a key predictor of adoption, it only partially explains this disparity, suggesting that other barriers are at play. Overall, our data provide granular insights into the multipurpose usage of GenAI, highlighting the dual need for targeted educational initiatives and further investigation into the underlying barriers to equitable participation that competence alone cannot explain.

Paperid: 586, https://arxiv.org/pdf/2512.01816.pdf

Abstract:
Current multimodal models aim to transcend the limitations of single-modality representations by unifying understanding and generation, often using text-to-image (T2I) tasks to calibrate semantic consistency. However, their reliance on static, single-image generation in training and evaluation leads to overfitting to static pattern matching and semantic fusion, while fundamentally hindering their ability to model dynamic processes that unfold over time. To address these constraints, we propose Envision-a causal event progression benchmark for chained text-to-multi-image generation. Grounded in world knowledge and structured by spatiotemporal causality, it reorganizes existing evaluation dimensions and includes 1,000 four-stage prompts spanning six scientific and humanities domains. To transition evaluation from single images to sequential frames and assess whether models truly internalize world knowledge while adhering to causal-temporal constraints, we introduce Envision-Score, a holistic metric integrating multi-dimensional consistency, physicality, and aesthetics. Comprehensive evaluation of 15 models (10 specialized T2I models, 5 unified models) uncovers: specialized T2I models demonstrate proficiency in aesthetic rendering yet lack intrinsic world knowledge. Unified multimodal models bridge this gap, consistently outperforming specialized counterparts in causal narrative coherence. However, even these unified architectures remain subordinate to closed-source models and struggle to overcome the core challenge of spatiotemporal consistency. This demonstrates that a focus on causally-isolated single images impedes multi-frame reasoning and generation, promoting static pattern matching over dynamic world modeling-ultimately limiting world knowledge internalization, generation.

Paperid: 587, https://arxiv.org/pdf/2512.00812.pdf

Abstract:
Multi-label classification (MLC) remains vulnerable to label imbalance, spurious correlations, and distribution shifts, challenges that are particularly detrimental to rare label prediction. To address these limitations, we introduce the Causal Cooperative Game (CCG) framework, which conceptualizes MLC as a cooperative multi-player interaction. CCG unifies explicit causal discovery via Neural Structural Equation Models with a counterfactual curiosity reward to drive robust feature learning. Furthermore, it incorporates a causal invariance loss to ensure generalization across diverse environments, complemented by a specialized enhancement strategy for rare labels. Extensive benchmarking demonstrates that CCG substantially outperforms strong baselines in both rare label prediction and overall robustness. Through rigorous ablation studies and qualitative analysis, we validate the efficacy and interpretability of our components, underscoring the potential of synergizing causal inference with cooperative game theory for advancing multi-label learning.

Paperid: 588, https://arxiv.org/pdf/2512.00582.pdf

Abstract:
Satire, a form of artistic expression combining humor with implicit critique, holds significant social value by illuminating societal issues. Despite its cultural and societal significance, satire comprehension, particularly in purely visual forms, remains a challenging task for current vision-language models. This task requires not only detecting satire but also deciphering its nuanced meaning and identifying the implicated entities. Existing models often fail to effectively integrate local entity relationships with global context, leading to misinterpretation, comprehension biases, and hallucinations. To address these limitations, we propose SatireDecoder, a training-free framework designed to enhance satirical image comprehension. Our approach proposes a multi-agent system performing visual cascaded decoupling to decompose images into fine-grained local and global semantic representations. In addition, we introduce a chain-of-thought reasoning strategy guided by uncertainty analysis, which breaks down the complex satire comprehension process into sequential subtasks with minimized uncertainty. Our method significantly improves interpretive accuracy while reducing hallucinations. Experimental results validate that SatireDecoder outperforms existing baselines in comprehending visual satire, offering a promising direction for vision-language reasoning in nuanced, high-level semantic tasks.

Paperid: 589, https://arxiv.org/pdf/2512.05472.pdf

Abstract:
Spiking Neural Networks (SNNs) are considered naturally suited for temporal processing, with membrane potential propagation widely regarded as the core temporal modeling mechanism. However, existing research lack analysis of its actual contributions in complex temporal tasks. We design Non-Stateful (NS) models progressively removing membrane propagation to quantify its stage-wise role. Experiments reveal a counterintuitive phenomenon: moderate removal in shallow or deep layers improves performance, while excessive removal causes collapse. We attribute this to spatio-temporal resource competition where neurons encode both semantics and dynamics within limited range, with temporal state consuming capacity for spatial learning. Based on this, we propose Spatial-Temporal Separable Network (STSep), decoupling residual blocks into independent spatial and temporal branches. The spatial branch focuses on semantic extraction while the temporal branch captures motion through explicit temporal differences. Experiments on Something-Something V2, UCF101, and HMDB51 show STSep achieves superior performance, with retrieval task and attention analysis confirming focus on motion rather than static appearance. This work provides new perspectives on SNNs' temporal mechanisms and an effective solution for spatiotemporal modeling in video understanding.

Paperid: 590, https://arxiv.org/pdf/2512.05461.pdf

Abstract:
Large language models (LLMs) are rapidly being integrated into computational social science research, yet their blackboxed training and designed stochastic elements in inference pose unique challenges for scientific inquiry. This article argues that applying LLMs to social scientific tasks requires explicit assessment of uncertainty-an expectation long established in both quantitative methodology in the social sciences and machine learning. We introduce a unified framework for evaluating LLM uncertainty along two dimensions: the task type (T), which distinguishes between classification, short-form, and long-form generation, and the validation type (V), which captures the availability of reference data or evaluative criteria. Drawing from both computer science and social science literature, we map existing uncertainty quantification (UQ) methods to this T-V typology and offer practical recommendations for researchers. Our framework provides both a methodological safeguard and a practical guide for integrating LLMs into rigorous social science research.

Paperid: 591, https://arxiv.org/pdf/2512.04459.pdf

Abstract:
The autonomous driving community is increasingly focused on addressing the challenges posed by out-of-distribution (OOD) driving scenarios. A dominant research trend seeks to enhance end-to-end (E2E) driving systems by integrating vision-language models (VLMs), leveraging their rich world knowledge and reasoning abilities to improve generalization across diverse environments. However, most existing VLMs or vision-language agents (VLAs) are built upon autoregressive (AR) models. In this paper, we observe that existing AR-based VLMs -- limited by causal attention and sequential token generation -- often fail to maintain consistency and controllability between high-level reasoning and low-level planning. In contrast, recent discrete diffusion VLMs equipped with bidirectional attention exhibit superior controllability and reliability through iterative denoising. Building on these observations, we introduce dVLM-AD, a diffusion-based vision-language model that unifies perception, structured reasoning, and low-level planning for end-to-end driving. Evaluated on nuScenes and WOD-E2E, dVLM-AD yields more consistent reasoning-action pairs and achieves planning performance comparable to existing driving VLM/VLA systems despite a modest backbone, outperforming AR-based baselines with a 9 percent improvement in behavior-trajectory consistency and a 6 percent increase in RFS on long-tail WOD-E2E scenarios. These results suggest a controllable and reliable pathway for scalable end-to-end driving.

Paperid: 592, https://arxiv.org/pdf/2512.04019.pdf

Abstract:
Recent works have demonstrated the viability of utilizing over-fitted implicit neural representations (INRs) as alternatives to autoencoder-based models for neural video compression. Among these INR-based video codecs, Neural Video Representation Compression (NVRC) was the first to adopt a fully end-to-end compression framework that compresses INRs, achieving state-of-the-art performance. Moreover, some recently proposed lightweight INRs have shown comparable performance to their baseline codecs with computational complexity lower than 10kMACs/pixel. In this work, we extend NVRC toward lightweight representations, and propose NVRC-Lite, which incorporates two key changes. Firstly, we integrated multi-scale feature grids into our lightweight neural representation, and the use of higher resolution grids significantly improves the performance of INRs at low complexity. Secondly, we address the issue that existing INRs typically leverage autoregressive models for entropy coding: these are effective but impractical due to their slow coding speed. In this work, we propose an octree-based context model for entropy coding high-dimensional feature grids, which accelerates the entropy coding module of the model. Our experimental results demonstrate that NVRC-Lite outperforms C3, one of the best lightweight INR-based video codecs, with up to 21.03% and 23.06% BD-rate savings when measured in PSNR and MS-SSIM, respectively, while achieving 8.4x encoding and 2.5x decoding speedup. The implementation of NVRC-Lite will be made available.

Paperid: 593, https://arxiv.org/pdf/2512.03402.pdf

Abstract:
Low-rank adaptation (LoRA) is one of the most popular methods among parameter-efficient fine-tuning (PEFT) methods to adapt pre-trained large language models (LLMs) to specific downstream tasks. However, the model trained based on LoRA often has an unsatisfactory performance due to its low-rank assumption. In this paper, we propose a novel method called Dual LoRA to improve the performance by incorporating an inductive bias into the original LoRA. Specifically, we separate low-rank matrices into two groups: the magnitude group to control whether or not and how far we should update a parameter and the direction group to decide whether this parameter should move forward or backward, to better simulate the parameter updating process of the full fine-tuning based on gradient-based optimization algorithms. We show that this can be simply achieved by adding a ReLU function to the magnitude group and a sign function to the direction group. We conduct several experiments over a wide range of NLP tasks, including natural language generation (NLG), understanding (NLU), and commonsense reasoning datasets on GPT-2, RoBERTa, DeBERTa, and LLaMA-1/2/3 as baseline models. The results show that we consistently outperform LoRA and its state-of-the-art variants with the same number of trainable parameters.

Paperid: 594, https://arxiv.org/pdf/2512.02224.pdf

Abstract:
Recent works in video quality assessment (VQA) typically employ monolithic models that typically predict a single quality score for each test video. These approaches cannot provide diagnostic, interpretable feedback, offering little insight into why the video quality is degraded. Most of them are also specialized, format-specific metrics rather than truly ``generic" solutions, as they are designed to learn a compromised representation from disparate perceptual domains. To address these limitations, this paper proposes Unified-VQA, a framework that provides a single, unified quality model applicable to various distortion types within multiple video formats by recasting generic VQA as a Diagnostic Mixture-of-Experts (MoE) problem. Unified-VQA employs multiple ``perceptual experts'' dedicated to distinct perceptual domains. A novel multi-proxy expert training strategy is designed to optimize each expert using a ranking-inspired loss, guided by the most suitable proxy metric for its domain. We also integrated a diagnostic multi-task head into this framework to generate a global quality score and an interpretable multi-dimensional artifact vector, which is optimized using a weakly-supervised learning strategy, leveraging the known properties of the large-scale training database generated for this work. With static model parameters (without retraining or fine-tuning), Unified-VQA demonstrates consistent and superior performance compared to over 18 benchmark methods for both generic VQA and diagnostic artifact detection tasks across 17 databases containing diverse streaming artifacts in HD, UHD, HDR and HFR formats. This work represents an important step towards practical, actionable, and interpretable video quality assessment.

Paperid: 595, https://arxiv.org/pdf/2512.01993.pdf

Abstract:
Autonomous driving policies are typically trained via open-loop behavior cloning of human demonstrations. However, such policies suffer from covariate shift when deployed in closed loop, leading to compounding errors. We introduce Rollouts as Demonstrations (RoaD), a simple and efficient method to mitigate covariate shift by leveraging the policy's own closed-loop rollouts as additional training data. During rollout generation, RoaD incorporates expert guidance to bias trajectories toward high-quality behavior, producing informative yet realistic demonstrations for fine-tuning. This approach enables robust closed-loop adaptation with orders of magnitude less data than reinforcement learning, and avoids restrictive assumptions of prior closed-loop supervised fine-tuning (CL-SFT) methods, allowing broader applications domains including end-to-end driving. We demonstrate the effectiveness of RoaD on WOSAC, a large-scale traffic simulation benchmark, where it performs similar or better than the prior CL-SFT method; and in AlpaSim, a high-fidelity neural reconstruction-based simulator for end-to-end driving, where it improves driving score by 41\% and reduces collisions by 54\%.

Paperid: 596, https://arxiv.org/pdf/2512.01202.pdf

Abstract:
With the rapid advances in programmable materials, reconfigurable intelligent surfaces (RIS) have become a pivotal technology for future wireless communications. The simultaneous transmitting and reflecting reconfigurable intelligent surfaces (STAR-RIS) can both transmit and reflect signals, enabling comprehensive signal control and expanding application scenarios. This paper introduces an unmanned aerial vehicle (UAV) to further enhance system flexibility and proposes an optimization design for the spectrum efficiency of the STAR-RIS-UAV-assisted wireless communication system. We present a deep reinforcement learning (DRL) algorithm capable of iteratively optimizing beamforming, phase shifts, and UAV positioning to maximize the system's sum rate through continuous interactions with the environment. To improve exploration in deterministic policies, we introduce a stochastic perturbation factor, which enhances exploration capabilities. As exploration is strengthened, the algorithm's ability to accurately evaluate the state-action value function becomes critical. Thus, based on the deep deterministic policy gradient (DDPG) algorithm, we propose a convolution-augmented deep deterministic policy gradient (CA-DDPG) algorithm that balances exploration and evaluation to improve the system's sum rate. The simulation results demonstrate that the CA-DDPG algorithm effectively interacts with the environment, optimizing the beamforming matrix, phase shift matrix, and UAV location, thereby improving system capacity and achieving better performance than other algorithms.

Paperid: 597, https://arxiv.org/pdf/2512.00524.pdf

Abstract:
Hierarchical clustering is a fundamental machine-learning technique for grouping data points into dendrograms. However, existing hierarchical clustering methods encounter two primary challenges: 1) Most methods specify dendrograms without a global objective. 2) Graph-based methods often neglect the significance of graph structure, optimizing objectives on complete or static predefined graphs. In this work, we propose Hyperbolic Continuous Structural Entropy neural networks, namely HypCSE, for structure-enhanced continuous hierarchical clustering. Our key idea is to map data points in the hyperbolic space and minimize the relaxed continuous structural entropy (SE) on structure-enhanced graphs. Specifically, we encode graph vertices in hyperbolic space using hyperbolic graph neural networks and minimize approximate SE defined on graph embeddings. To make the SE objective differentiable for optimization, we reformulate it into a function using the lowest common ancestor (LCA) on trees and then relax it into continuous SE (CSE) by the analogy of hyperbolic graph embeddings and partitioning trees. To ensure a graph structure that effectively captures the hierarchy of data points for CSE calculation, we employ a graph structure learning (GSL) strategy that updates the graph structure during training. Extensive experiments on seven datasets demonstrate the superior performance of HypCSE.

Paperid: 598, https://arxiv.org/pdf/2512.05647.pdf

Abstract:
We introduce an open, machine-readable corpus of Greek government decisions sourced from the national transparency platform Diavgeia. The resource comprises 1 million decisions, featuring and high-quality raw text extracted from PDFs. It is released with raw extracted text in Markdown format, alongside a fully reproducible extraction pipeline. Beyond the core dataset, we conduct qualitative analyses to explore boilerplate patterns and design a retrieval-augmented generation (RAG) task by formulating a set of representative questions, creating high-quality answers, and evaluating a baseline RAG system on its ability to retrieve and reason over public decisions. This evaluation demonstrates the potential of large-scale public-sector corpora to support advanced information access and transparency through structured retrieval and reasoning over governmental documents, and highlights how such a RAG pipeline could simulate a chat-based assistant capable of interactively answering questions about public decisions. Due to its scale, quality, and domain coverage, the corpus can also serve as high-value pre-training or fine-tuning material for new Language Models (LMs) and Large Language Models (LLMs) respectively, including specialized models for legal and governmental domains, and as a foundation for novel approaches in domain adaptation, knowledge-grounded generation, and explainable AI. Finally, we discuss limitations, outline future directions, and make both the data and the code accessible.

Paperid: 599, https://arxiv.org/pdf/2512.03377.pdf

Abstract:
Transformers have achieved significant success across various domains, relying on self-attention to capture dependencies. However, the standard first-order attention mechanism is often limited by a low-rank bottleneck, struggling to capture intricate, multi-hop relationships within a single layer. In this paper, we propose the Nexus, a novel architecture designed to enhance representational power through a recursive framework. Unlike standard approaches that use static linear projections for Queries and Keys, Nexus dynamically refines these representations via nested self-attention mechanisms. Specifically, the Query and Key vectors are themselves outputs of inner attention loops, allowing tokens to aggregate global context and model high-order correlations \textit{prior} to the final attention computation. We enforce a parameter-efficient weight-sharing strategy across recursive steps, ensuring that this enhanced expressivity incurs $\mathcal{O}(1)$ additional parameters. We provide theoretical analysis demonstrating that our method breaks the linear bottleneck of standard attention. Empirically, Nexus outperforms standard Transformers on multiple benchmarks.

Paperid: 600, https://arxiv.org/pdf/2512.02550.pdf

Abstract:
The computational demands of modern Deep Neural Networks (DNNs) are immense and constantly growing. While training costs usually capture public attention, inference demands are also contributing in significant computational, energy and environmental footprints. Sparsity stands out as a critical mechanism for drastically reducing these resource demands. However, its potential remains largely untapped and is not yet fully incorporated in production AI systems. To bridge this gap, this work provides the necessary knowledge and insights for performance engineers keen to get involved in deep learning inference optimization. In particular, in this work we: a) discuss the various forms of sparsity that can be utilized in DNN inference, b) explain how the original dense computations translate to sparse kernels, c) provide an extensive bibliographic review of the state-of-the-art in the implementation of these kernels for CPUs and GPUs, d) discuss the availability of sparse datasets in support of sparsity-related research and development, e) explore the current software tools and frameworks that provide robust sparsity support, and f) present evaluation results of different implementations of the key SpMM and SDDMM kernels on CPU and GPU platforms. Ultimately, this paper aims to serve as a resource for performance engineers seeking to develop and deploy highly efficient sparse deep learning models in productions.

Paperid: 601, https://arxiv.org/pdf/2512.01559.pdf

Abstract:
This paper introduces LLM2Fx-Tools, a multimodal tool-calling framework that generates executable sequences of audio effects (Fx-chain) for music post-production. LLM2Fx-Tools uses a large language model (LLM) to understand audio inputs, select audio effects types, determine their order, and estimate parameters, guided by chain-of-thought (CoT) planning. We also present LP-Fx, a new instruction-following dataset with structured CoT annotations and tool calls for audio effects modules. Experiments show that LLM2Fx-Tools can infer an Fx-chain and its parameters from pairs of unprocessed and processed audio, enabled by autoregressive sequence modeling, tool calling, and CoT reasoning. We further validate the system in a style transfer setting, where audio effects information is transferred from a reference source and applied to new content. Finally, LLM-as-a-judge evaluation demonstrates that our approach generates appropriate CoT reasoning and responses for music production queries. To our knowledge, this is the first work to apply LLM-based tool calling to audio effects modules, enabling interpretable and controllable music production.

Paperid: 602, https://arxiv.org/pdf/2512.07246.pdf

Abstract:
Error detection (ED), which aims to identify incorrect or inconsistent cell values in tabular data, is important for ensuring data quality. Recent state-of-the-art ED methods leverage the pre-trained knowledge and semantic capability embedded in large language models (LLMs) to directly label whether a cell is erroneous. However, this LLM-as-a-labeler pipeline (1) relies on the black box, implicit decision process, thus failing to provide explainability for the detection results, and (2) is highly sensitive to prompts, yielding inconsistent outputs due to inherent model stochasticity, therefore lacking robustness. To address these limitations, we propose an LLM-as-an-inducer framework that adopts LLM to induce the decision tree for ED (termed TreeED) and further ensembles multiple such trees for consensus detection (termed ForestED), thereby improving explainability and robustness. Specifically, based on prompts derived from data context, decision tree specifications and output requirements, TreeED queries the LLM to induce the decision tree skeleton, whose root-to-leaf decision paths specify the stepwise procedure for evaluating a given sample. Each tree contains three types of nodes: (1) rule nodes that perform simple validation checks (e.g., format or range), (2) Graph Neural Network (GNN) nodes that capture complex patterns (e.g., functional dependencies), and (3) leaf nodes that output the final decision types (error or clean). Furthermore, ForestED employs uncertainty-based sampling to obtain multiple row subsets, constructing a decision tree for each subset using TreeED. It then leverages an Expectation-Maximization-based algorithm that jointly estimates tree reliability and optimizes the consensus ED prediction. Extensive xperiments demonstrate that our methods are accurate, explainable and robust, achieving an average F1-score improvement of 16.1% over the best baseline.

Paperid: 603, https://arxiv.org/pdf/2512.05470.pdf

Abstract:
Generative AI (GenAI) has reshaped software system design by introducing foundation models as pre-trained subsystems that redefine architectures and operations. The emerging challenge is no longer model fine-tuning but context engineering-how systems capture, structure, and govern external knowledge, memory, tools, and human input to enable trustworthy reasoning. Existing practices such as prompt engineering, retrieval-augmented generation (RAG), and tool integration remain fragmented, producing transient artefacts that limit traceability and accountability. This paper proposes a file-system abstraction for context engineering, inspired by the Unix notion that 'everything is a file'. The abstraction offers a persistent, governed infrastructure for managing heterogeneous context artefacts through uniform mounting, metadata, and access control. Implemented within the open-source AIGNE framework, the architecture realises a verifiable context-engineering pipeline, comprising the Context Constructor, Loader, and Evaluator, that assembles, delivers, and validates context under token constraints. As GenAI becomes an active collaborator in decision support, humans play a central role as curators, verifiers, and co-reasoners. The proposed architecture establishes a reusable foundation for accountable and human-centred AI co-work, demonstrated through two exemplars: an agent with memory and an MCP-based GitHub assistant. The implementation within the AIGNE framework demonstrates how the architecture can be operationalised in developer and industrial settings, supporting verifiable, maintainable, and industry-ready GenAI systems.

Paperid: 604, https://arxiv.org/pdf/2512.02161.pdf

Abstract:
Text-to-image (T2I) models are capable of generating visually impressive images, yet they often fail to accurately capture specific attributes in user prompts, such as the correct number of objects with the specified colors. The diversity of such errors underscores the need for a hierarchical evaluation framework that can compare prompt adherence abilities of different image generation models. Simultaneously, benchmarks of vision language models (VLMs) have not kept pace with the complexity of scenes that VLMs are used to annotate. In this work, we propose a structured methodology for jointly evaluating T2I models and VLMs by testing whether VLMs can identify 27 specific failure modes in the images generated by T2I models conditioned on challenging prompts. Our second contribution is a dataset of prompts and images generated by 5 T2I models (Flux, SD3-Medium, SD3-Large, SD3.5-Medium, SD3.5-Large) and the corresponding annotations from VLMs (Molmo, InternVL3, Pixtral) annotated by an LLM (Llama3) to test whether VLMs correctly identify the failure mode in a generated image. By analyzing failure modes on a curated set of prompts, we reveal systematic errors in attribute fidelity and object representation. Our findings suggest that current metrics are insufficient to capture these nuanced errors, highlighting the importance of targeted benchmarks for advancing generative model reliability and interpretability.

Paperid: 605, https://arxiv.org/pdf/2512.01925.pdf

Abstract:
Recent advancements in large language models (LLMs) have been driven by their emergent reasoning capabilities, particularly through long chain-of-thought (CoT) prompting, which enables thorough exploration and deliberation. Despite these advances, long-CoT LLMs often exhibit suboptimal reasoning behaviors, such as overthinking and excessively protracted reasoning chains, which can impair performance. In this paper, we analyze reasoning processes through an optimization lens, framing CoT as a gradient descent procedure where each reasoning step constitutes an update toward problem resolution. Building on this perspective, we introduce RePro (Rectifying Process-level Reward), a novel approach to refine LLM reasoning during post-training. RePro defines a surrogate objective function to assess the optimization process underlying CoT, utilizing a dual scoring mechanism to quantify its intensity and stability. These scores are aggregated into a composite process-level reward, seamlessly integrated into reinforcement learning with verifiable rewards (RLVR) pipelines to optimize LLMs. Extensive experiments across multiple reinforcement learning algorithms and diverse LLMs, evaluated on benchmarks spanning mathematics, science, and coding, demonstrate that RePro consistently enhances reasoning performance and mitigates suboptimal reasoning behaviors.

Paperid: 606, https://arxiv.org/pdf/2512.07209.pdf

Abstract:
We introduce a novel pipeline for joint audio-visual editing that enhances the coherence between edited video and its accompanying audio. Our approach first applies state-of-the-art video editing techniques to produce the target video, then performs audio editing to align with the visual changes. To achieve this, we present a new video-to-audio generation model that conditions on the source audio, target video, and a text prompt. We extend the model architecture to incorporate conditional audio input and propose a data augmentation strategy that improves training efficiency. Furthermore, our model dynamically adjusts the influence of the source audio based on the complexity of the edits, preserving the original audio structure where possible. Experimental results demonstrate that our method outperforms existing approaches in maintaining audio-visual alignment and content integrity.

Paperid: 607, https://arxiv.org/pdf/2512.02982.pdf

Abstract:
Modeling dynamic 3D environments from LiDAR sequences is central to building reliable 4D worlds for autonomous driving and embodied AI. Existing generative frameworks, however, often treat all spatial regions uniformly, overlooking the varying uncertainty across real-world scenes. This uniform generation leads to artifacts in complex or ambiguous regions, limiting realism and temporal stability. In this work, we present U4D, an uncertainty-aware framework for 4D LiDAR world modeling. Our approach first estimates spatial uncertainty maps from a pretrained segmentation model to localize semantically challenging regions. It then performs generation in a "hard-to-easy" manner through two sequential stages: (1) uncertainty-region modeling, which reconstructs high-entropy regions with fine geometric fidelity, and (2) uncertainty-conditioned completion, which synthesizes the remaining areas under learned structural priors. To further ensure temporal coherence, U4D incorporates a mixture of spatio-temporal (MoST) block that adaptively fuses spatial and temporal representations during diffusion. Extensive experiments show that U4D produces geometrically faithful and temporally consistent LiDAR sequences, advancing the reliability of 4D world modeling for autonomous perception and simulation.

Paperid: 608, https://arxiv.org/pdf/2512.07287.pdf

Abstract:
Despite impressive advances in agent systems, multi-turn tool-use scenarios remain challenging. It is mainly because intent is clarified progressively and the environment evolves with each tool call. While reusing past experience is natural, current LLM agents either treat entire trajectories or pre-defined subtasks as indivisible units, or solely exploit tool-to-tool dependencies, hindering adaptation as states and information evolve across turns. In this paper, we propose a State Integrated Tool Graph (SIT-Graph), which enhances multi-turn tool use by exploiting partially overlapping experience. Inspired by human decision-making that integrates episodic and procedural memory, SIT-Graph captures both compact state representations (episodic-like fragments) and tool-to-tool dependencies (procedural-like routines) from historical trajectories. Specifically, we first build a tool graph from accumulated tool-use sequences, and then augment each edge with a compact state summary of the dialog and tool history that may shape the next action. At inference time, SIT-Graph enables a human-like balance between episodic recall and procedural execution: when the next decision requires recalling prior context, the agent retrieves the state summaries stored on relevant edges and uses them to guide its next action; when the step is routine, it follows high-confidence tool dependencies without explicit recall. Experiments across multiple stateful multi-turn tool-use benchmarks show that SIT-Graph consistently outperforms strong memory- and graph-based baselines, delivering more robust tool selection and more effective experience transfer.

Paperid: 609, https://arxiv.org/pdf/2512.07212.pdf

Abstract:
Imitation learning with diffusion models has advanced robotic control by capturing multi-modal action distributions. However, existing approaches typically treat observations as high-level conditioning inputs to the denoising network, rather than integrating them into the stochastic dynamics of the diffusion process itself. As a result, sampling must begin from random Gaussian noise, weakening the coupling between perception and control and often yielding suboptimal performance. We introduce BridgePolicy, a generative visuomotor policy that explicitly embeds observations within the stochastic differential equation via a diffusion-bridge formulation. By constructing an observation-informed trajectory, BridgePolicy enables sampling to start from a rich, informative prior rather than random noise, substantially improving precision and reliability in control. A key challenge is that classical diffusion bridges connect distributions with matched dimensionality, whereas robotic observations are heterogeneous and multi-modal and do not naturally align with the action space. To address this, we design a multi-modal fusion module and a semantic aligner that unify visual and state inputs and align observation and action representations, making the bridge applicable to heterogeneous robot data. Extensive experiments across 52 simulation tasks on three benchmarks and five real-world tasks demonstrate that BridgePolicy consistently outperforms state-of-the-art generative policies.

Paperid: 610, https://arxiv.org/pdf/2512.05651.pdf

Abstract:
The proliferation of AI-generated imagery poses escalating challenges for multimedia forensics, yet many existing detectors depend on assumptions about the internals of specific generative models, limiting their cross-model applicability. We introduce a self-supervised approach for detecting AI-generated images that leverages camera metadata -- specifically exchangeable image file format (EXIF) tags -- to learn features intrinsic to digital photography. Our pretext task trains a feature extractor solely on camera-captured photographs by classifying categorical EXIF tags (\eg, camera model and scene type) and pairwise-ranking ordinal and continuous EXIF tags (\eg, focal length and aperture value). Using these EXIF-induced features, we first perform one-class detection by modeling the distribution of photographic images with a Gaussian mixture model and flagging low-likelihood samples as AI-generated. We then extend to binary detection that treats the learned extractor as a strong regularizer for a classifier of the same architecture, operating on high-frequency residuals from spatially scrambled patches. Extensive experiments across various generative models demonstrate that our EXIF-induced detectors substantially advance the state of the art, delivering strong generalization to in-the-wild samples and robustness to common benign image perturbations.

Paperid: 611, https://arxiv.org/pdf/2512.04585.pdf

Abstract:
Segment Anything Model 3 (SAM3) has advanced open-vocabulary segmentation through promptable concept segmentation, allowing users to segment all instances corresponding to a given concept, typically specified with short noun-phrase (NP) prompts. While this marks the first integration of language-level concepts within the SAM family, real-world usage typically requires far richer expressions that include attributes, spatial relations, functionalities, actions, states, and even implicit reasoning over instances. Currently, SAM3 relies on external multi-modal agents to convert complex instructions into NPs and then conduct iterative mask filtering. However, these NP-level concepts remain overly coarse, often failing to precisely represent a specific instance. In this work, we present SAM3-I, an enhanced framework that unifies concept-level understanding and instruction-level reasoning within the SAM family. SAM3-I introduces an instruction-aware cascaded adaptation mechanism that progressively aligns expressive instruction semantics with SAM3's existing vision-language representations, enabling direct instruction-following segmentation without sacrificing its original concept-driven capabilities. Furthermore, we design a structured instruction taxonomy spanning concept, simple, and complex levels, and develop a scalable data engine to construct a dataset with diverse instruction-mask pairs. Experiments show that SAM3-I delivers appealing performance, demonstrating that SAM3 can be effectively extended to follow natural-language instructions while preserving its strong concept grounding. We open-source SAM3-I and provide practical fine-tuning workflows, enabling researchers to adapt it to domain-specific applications. The source code is available here.

Paperid: 612, https://arxiv.org/pdf/2512.04535.pdf

Abstract:
The integration of external tools is pivotal for empowering Large Language Model (LLM) agents with real-world capabilities. However, training these agents through direct, continuous interaction with diverse tools is often prohibitively expensive, slow, and introduces additional development and maintenance overhead. To address this challenge, we introduce the Generalist Tool Model (GTM), a 1.5-billion-parameter model that learns to act as a universal tool simulator. With only prompt-level configuration, GTM accesses tool functionalities along with input arguments and generates outputs that faithfully mimic real tool execution, providing a fast and cost-effective solution that eliminates development overhead. To build GTM, we propose the Context-Aware Response Generation (CARG) pipeline, which synthesizes comprehensive training data covering over 20,000 tools across 300 domains including physics, medicine, robotics, and finance. Through this pipeline, GTM learns to produce not only syntactically correct outputs but also logically coherent and contextually appropriate responses. Experiments demonstrate that GTM produces high-quality outputs with strong consistency and reliability. Besides when used in real reinforcement learning scenarios for agent training, GTM exhibits significantly faster simulation speed compared to real tools while maintaining comparable output quality, along with remarkable generalization and domain adaptability. Our results establish GTM as a foundational component for developing future AI agents, enabling efficient and scalable training of tool-augmented systems.

Paperid: 613, https://arxiv.org/pdf/2512.02536.pdf

Abstract:
Recent progress in multimodal large language models (MLLMs) has highlighted the challenge of efficiently bridging pre-trained Vision-Language Models (VLMs) with Diffusion Models. While methods using a fixed number of learnable query tokens offer computational efficiency, they suffer from task generalization collapse, failing to adapt to new tasks that are distant from their pre-training tasks. To overcome this, we propose Noisy Query Tokens, which learn a distributed representation space between the VLM and Diffusion Model via end-to-end optimization, enhancing continual learning. Additionally, we introduce a VAE branch with linear projection to recover fine-grained image details. Experimental results confirm our approach mitigates generalization collapse and enables stable continual learning across diverse tasks.

Paperid: 614, https://arxiv.org/pdf/2512.00725.pdf

Abstract:
Typical deep clustering methods, while achieving notable progress, can only provide one clustering result per dataset. This limitation arises from their assumption of a fixed underlying data distribution, which may fail to meet user needs and provide unsatisfactory clustering outcomes. Our work investigates how multi-modal large language models (MLLMs) can be leveraged to achieve user-driven clustering, emphasizing their adaptability to user-specified semantic requirements. However, directly using MLLM output for clustering has risks for producing unstructured and generic image descriptions instead of feature-specific and concrete ones. To address these issues, our method first discovers that MLLMs' hidden states of text tokens are strongly related to the corresponding features, and leverages these embeddings to perform clusterings from any user-defined criteria. We also employ a lightweight clustering head augmented with pseudo-label learning, significantly enhancing clustering accuracy. Extensive experiments demonstrate its competitive performance on diverse datasets and metrics.

Paperid: 615, https://arxiv.org/pdf/2512.00097.pdf

Abstract:
Automated theorem proving in Euclidean geometry, particularly for International Mathematical Olympiad (IMO) level problems, remains a major challenge and an important research focus in Artificial Intelligence. In this paper, we present a highly efficient method for geometry theorem proving that runs entirely on CPUs without relying on neural network-based inference. Our initial study shows that a simple random strategy for adding auxiliary points can achieve silver-medal level human performance on IMO. Building on this, we propose HAGeo, a Heuristic-based method for adding Auxiliary constructions in Geometric deduction that solves 28 of 30 problems on the IMO-30 benchmark, achieving gold-medal level performance and surpassing AlphaGeometry, a competitive neural network-based approach, by a notable margin. To evaluate our method and existing approaches more comprehensively, we further construct HAGeo-409, a benchmark consisting of 409 geometry problems with human-assessed difficulty levels. Compared with the widely used IMO-30, our benchmark poses greater challenges and provides a more precise evaluation, setting a higher bar for geometry theorem proving.

Paperid: 616, https://arxiv.org/pdf/2512.05085.pdf

Abstract:
This paper investigates the impact of the recently proposed concept of fluid reconfigurable intelligent surfaces (FRIS) on covert communications. Specifically, we consider a communication scenario where a legitimate transmitter aims to covertly deliver information to its intended receiver through a planar FRIS, while an adversary attempts to detect whether any transmission is occurring. In this context, we analyze the false alarm (FA) and missed detection (MD) probabilities, and derive a closed-form expression for the covertness outage probability (COP). Furthermore, the success probability is characterized under the optimal detection threshold, providing new insights into the trade-off between covertness and reliable transmission. Numerical results reveal that FRIS provides a clear advantage over fixed-position RIS at low-to-moderate transmit powers by improving reliability and enhancing covertness, while at very high power levels, fixed-position RIS may sustain slightly higher success probability due to reduced leakage toward the adversary.

Paperid: 617, https://arxiv.org/pdf/2512.04554.pdf

Abstract:
Document Visual Question Answering (DocVQA) enables end-to-end reasoning grounded on information present in a document input. While recent models have shown impressive capabilities, they remain vulnerable to adversarial attacks. In this work, we introduce a novel attack scenario that aims to forge document content in a visually imperceptible yet semantically targeted manner, allowing an adversary to induce specific or generally incorrect answers from a DocVQA model. We develop specialized attack algorithms that can produce adversarially forged documents tailored to different attackers' goals, ranging from targeted misinformation to systematic model failure scenarios. We demonstrate the effectiveness of our approach against two end-to-end state-of-the-art models: Pix2Struct, a vision-language transformer that jointly processes image and text through sequence-to-sequence modeling, and Donut, a transformer-based model that directly extracts text and answers questions from document images. Our findings highlight critical vulnerabilities in current DocVQA systems and call for the development of more robust defenses.

Paperid: 618, https://arxiv.org/pdf/2512.04359.pdf

Abstract:
Reinforcement learning with verifiable rewards (RLVR) has demonstrated superior performance in enhancing the reasoning capability of large language models (LLMs). However, this accuracy-oriented learning paradigm often suffers from entropy collapse, which reduces policy exploration and limits reasoning capabilities. To address this challenge, we propose an efficient reinforcement learning framework that leverages entropy signals at both the semantic and token levels to improve reasoning. From the data perspective, we introduce semantic entropy-guided curriculum learning, organizing training data from low to high semantic entropy to guide progressive optimization from easier to more challenging tasks. For the algorithmic design, we adopt non-uniform token treatment by imposing KL regularization on low-entropy tokens that critically impact policy exploration and applying stronger constraints on high-covariance portions within these tokens. By jointly optimizing data organization and algorithmic design, our method effectively mitigates entropy collapse and enhances LLM reasoning. Experimental results across 6 benchmarks with 3 different parameter-scale base models demonstrate that our method outperforms other entropy-based approaches in improving reasoning.

Paperid: 619, https://arxiv.org/pdf/2512.03528.pdf

Abstract:
Communication is one of the effective means to improve the learning of cooperative policy in multi-agent systems. However, in most real-world scenarios, lossy communication is a prevalent issue. Existing multi-agent reinforcement learning with communication, due to their limited scalability and robustness, struggles to apply to complex and dynamic real-world environments. To address these challenges, we propose a generalized communication-constrained model to uniformly characterize communication conditions across different scenarios. Based on this, we utilize it as a learning prior to distinguish between lossy and lossless messages for specific scenarios. Additionally, we decouple the impact of lossy and lossless messages on distributed decision-making, drawing on a dual mutual information estimatior, and introduce a communication-constrained multi-agent reinforcement learning framework, quantifying the impact of communication messages into the global reward. Finally, we validate the effectiveness of our approach across several communication-constrained benchmarks.

Paperid: 620, https://arxiv.org/pdf/2512.02993.pdf

Abstract:
Prevailing 3D texture generation methods, which often rely on multi-view fusion, are frequently hindered by inter-view inconsistencies and incomplete coverage of complex surfaces, limiting the fidelity and completeness of the generated content. To overcome these challenges, we introduce TEXTRIX, a native 3D attribute generation framework for high-fidelity texture synthesis and downstream applications such as precise 3D part segmentation. Our approach constructs a latent 3D attribute grid and leverages a Diffusion Transformer equipped with sparse attention, enabling direct coloring of 3D models in volumetric space and fundamentally avoiding the limitations of multi-view fusion. Built upon this native representation, the framework naturally extends to high-precision 3D segmentation by training the same architecture to predict semantic attributes on the grid. Extensive experiments demonstrate state-of-the-art performance on both tasks, producing seamless, high-fidelity textures and accurate 3D part segmentation with precise boundaries.

Paperid: 621, https://arxiv.org/pdf/2512.01948.pdf

Abstract:
Deep Research Agents (DRAs) aim to automatically produce analyst-level reports through iterative information retrieval and synthesis. However, most existing DRAs were validated on question-answering benchmarks, while research on generating comprehensive reports remains overlooked. Worse, current benchmarks for report synthesis suffer from task complexity and subjective metrics -- this fails to reflect user demands and limits the practical utility of generated reports. To address these gaps, we present Fine-grained DEepResearch bench (FINDER), an enhanced benchmark consisting of 100 human-curated research tasks with 419 structured checklist items that standardize report structure, analytical depth, and factual grounding. Based on approximately 1,000 reports produced by mainstream DRAs, we further propose Deep rEsearch Failure Taxonomy (DEFT), the first failure taxonomy for deep research agents. DEFT contains 14 fine-grained failure modes across reasoning, retrieval, and generation, and is built upon grounded theory with human-LLM co-annotating and inter-annotator reliability validation. Our experimental findings reveal that current DRAs struggle not with task comprehension but with evidence integration, verification, and reasoning-resilient planning.

Paperid: 622, https://arxiv.org/pdf/2512.01374.pdf

Abstract:
This paper proposes a novel formulation for reinforcement learning (RL) with large language models, explaining why and under what conditions the true sequence-level reward can be optimized via a surrogate token-level objective in policy gradient methods such as REINFORCE. Specifically, through a first-order approximation, we show that this surrogate becomes increasingly valid only when both the training-inference discrepancy and policy staleness are minimized. This insight provides a principled explanation for the crucial role of several widely adopted techniques in stabilizing RL training, including importance sampling correction, clipping, and particularly Routing Replay for Mixture-of-Experts (MoE) models. Through extensive experiments with a 30B MoE model totaling hundreds of thousands of GPU hours, we show that for on-policy training, the basic policy gradient algorithm with importance sampling correction achieves the highest training stability. When off-policy updates are introduced to accelerate convergence, combining clipping and Routing Replay becomes essential to mitigate the instability caused by policy staleness. Notably, once training is stabilized, prolonged optimization consistently yields comparable final performance regardless of cold-start initialization. We hope that the shared insights and the developed recipes for stable RL training will facilitate future research.

Paperid: 623, https://arxiv.org/pdf/2512.04763.pdf

Abstract:
Memory-augmented Large Language Models (LLMs) have demonstrated remarkable consistency during prolonged dialogues by storing relevant memories and incorporating them as context. Such memory-based personalization is also key in on-device settings that allow users to keep their conversations and data private. However, memory-augmented systems typically rely on LLMs that are too costly for local on-device deployment. Even though Small Language Models (SLMs) are more suitable for on-device inference than LLMs, they cannot achieve sufficient performance. Additionally, these LLM-based systems lack native visual capabilities, limiting their applicability in multimodal contexts. In this paper, we introduce (i) MemLoRA, a novel memory system that enables local deployment by equipping SLMs with specialized memory adapters, and (ii) its vision extension MemLoRA-V, which integrates small Vision-Language Models (SVLMs) to memory systems, enabling native visual understanding. Following knowledge distillation principles, each adapter is trained separately for specific memory operations$\unicode{x2013}$knowledge extraction, memory update, and memory-augmented generation. Equipped with memory adapters, small models enable accurate on-device memory operations without cloud dependency. On text-only operations, MemLoRA outperforms 10$\times$ larger baseline models (e.g., Gemma2-27B) and achieves performance comparable to 60$\times$ larger models (e.g., GPT-OSS-120B) on the LoCoMo benchmark. To evaluate visual understanding operations instead, we extend LoCoMo with challenging Visual Question Answering tasks that require direct visual reasoning. On this, our VLM-integrated MemLoRA-V shows massive improvements over caption-based approaches (81.3 vs. 23.7 accuracy) while keeping strong performance in text-based tasks, demonstrating the efficacy of our method in multimodal contexts.

Paperid: 624, https://arxiv.org/pdf/2512.01665.pdf

Abstract:
Tiny object detection in remote sensing imagery has attracted significant research interest in recent years. Despite recent progress, achieving balanced detection performance across diverse object scales remains a formidable challenge, particularly in scenarios where dense tiny objects and large objects coexist. Although large foundation models have revolutionized general vision tasks, their application to tiny object detection remains unexplored due to the extreme scale variation and density distribution inherent to remote sensing imagery. To bridge this scale gap, we propose ScaleBridge-Det, to the best of our knowledge, the first large detection framework designed for tiny objects, which could achieve balanced performance across diverse scales through scale-adaptive expert routing and density-guided query allocation. Specifically, we introduce a Routing-Enhanced Mixture Attention (REM) module that dynamically selects and fuses scale-specific expert features via adaptive routing to address the tendency of standard MoE models to favor dominant scales. REM generates complementary and discriminative multi-scale representations suitable for both tiny and large objects. Furthermore, we present a Density-Guided Dynamic Query (DGQ) module that predicts object density to adaptively adjust query positions and numbers, enabling efficient resource allocation for objects of varying scales. The proposed framework allows ScaleBridge-Det to simultaneously optimize performance for both dense tiny and general objects without trade-offs. Extensive experiments on benchmark and cross-domain datasets demonstrate that ScaleBridge-Det achieves state-of-the-art performance on AI-TOD-V2 and DTOD, while exhibiting superior cross-domain robustness on VisDrone.

Paperid: 625, https://arxiv.org/pdf/2512.01095.pdf

Abstract:
We present CycliST, a novel benchmark dataset designed to evaluate Video Language Models (VLM) on their ability for textual reasoning over cyclical state transitions. CycliST captures fundamental aspects of real-world processes by generating synthetic, richly structured video sequences featuring periodic patterns in object motion and visual attributes. CycliST employs a tiered evaluation system that progressively increases difficulty through variations in the number of cyclic objects, scene clutter, and lighting conditions, challenging state-of-the-art models on their spatio-temporal cognition. We conduct extensive experiments with current state-of-the-art VLMs, both open-source and proprietary, and reveal their limitations in generalizing to cyclical dynamics such as linear and orbital motion, as well as time-dependent changes in visual attributes like color and scale. Our results demonstrate that present-day VLMs struggle to reliably detect and exploit cyclic patterns, lack a notion of temporal understanding, and are unable to extract quantitative insights from scenes, such as the number of objects in motion, highlighting a significant technical gap that needs to be addressed. More specifically, we find no single model consistently leads in performance: neither size nor architecture correlates strongly with outcomes, and no model succeeds equally well across all tasks. By providing a targeted challenge and a comprehensive evaluation framework, CycliST paves the way for visual reasoning models that surpass the state-of-the-art in understanding periodic patterns.

Paperid: 626, https://arxiv.org/pdf/2512.00847.pdf

Abstract:
Diffractive neural networks, where signal processing is embedded into wave propagation, promise light-speed and energy-efficient computation. However, existing three-dimensional structures, such as stacked intelligent metasurfaces (SIMs), face critical challenges in implementation and integration. In contrast, this work pioneers planar diffractive neural networks (PDNNs) empowered communications, a novel architecture that performs signal processing as signals propagate through artificially designed planar circuits. To demonstrate the capability of PDNN, we propose a PDNN-based space-shift-keying (PDNN-SSK) communication system with a single radio-frequency (RF) chain and a maximum power detector. In this system, PDNNs are deployed at both the transmitter and receiver to jointly execute modulation, beamforming, and detection. We conduct theoretical analyses to provide the maximization condition of correct detection probability and derive the closed-form expression of the symbol error rate (SER) for the proposed system. To approach these theoretical benchmarks, the phase shift parameters of PDNNs are optimized using a surrogate model-based training approach, which effectively navigates the high-dimensional, non-convex optimization landscape. Extensive simulations verify the theoretical analysis framework and uncover fundamental design principles for the PDNN architecture, highlighting its potential to revolutionize RF front-ends by replacing conventional digital baseband modules with this integrable RF computing platform.

Paperid: 627, https://arxiv.org/pdf/2512.02973.pdf

Abstract:
While Multimodal Large Language Models (MLLMs) show remarkable capabilities, their safety alignments are susceptible to jailbreak attacks. Existing attack methods typically focus on text-image interplay, treating the visual modality as a secondary prompt. This approach underutilizes the unique potential of images to carry complex, contextual information. To address this gap, we propose a new image-centric attack method, Contextual Image Attack (CIA), which employs a multi-agent system to subtly embeds harmful queries into seemingly benign visual contexts using four distinct visualization strategies. To further enhance the attack's efficacy, the system incorporate contextual element enhancement and automatic toxicity obfuscation techniques. Experimental results on the MMSafetyBench-tiny dataset show that CIA achieves high toxicity scores of 4.73 and 4.83 against the GPT-4o and Qwen2.5-VL-72B models, respectively, with Attack Success Rates (ASR) reaching 86.31\% and 91.07\%. Our method significantly outperforms prior work, demonstrating that the visual modality itself is a potent vector for jailbreaking advanced MLLMs.

Paperid: 628, https://arxiv.org/pdf/2512.02851.pdf

Abstract:
Visual traversability estimation is critical for autonomous navigation, but existing VLM-based methods rely on hand-crafted prompts, generalize poorly across embodiments, and output only traversability maps, leaving trajectory generation to slow external planners. We propose SwarmDiffusion, a lightweight end-to-end diffusion model that jointly predicts traversability and generates a feasible trajectory from a single RGB image. To remove the need for annotated or planner-produced paths, we introduce a planner-free trajectory construction pipeline based on randomized waypoint sampling, Bezier smoothing, and regularization enforcing connectivity, safety, directionality, and path thinness. This enables learning stable motion priors without demonstrations. SwarmDiffusion leverages VLM-derived supervision without prompt engineering and conditions the diffusion process on a compact embodiment state, producing physically consistent, traversable paths that transfer across different robot platforms. Across indoor environments and two embodiments (quadruped and aerial), the method achieves 80-100% navigation success and 0.09s inference, and adapts to a new robot using only-500 additional visual samples. It generalizes reliably to unseen environments in simulation and real-world trials, offering a scalable, prompt-free approach to unified traversability reasoning and trajectory generation.

Paperid: 629, https://arxiv.org/pdf/2512.02038.pdf

Abstract:
Large language models (LLMs) have rapidly evolved from text generators into powerful problem solvers. Yet, many open tasks demand critical thinking, multi-source, and verifiable outputs, which are beyond single-shot prompting or standard retrieval-augmented generation. Recently, numerous studies have explored Deep Research (DR), which aims to combine the reasoning capabilities of LLMs with external tools, such as search engines, thereby empowering LLMs to act as research agents capable of completing complex, open-ended tasks. This survey presents a comprehensive and systematic overview of deep research systems, including a clear roadmap, foundational components, practical implementation techniques, important challenges, and future directions. Specifically, our main contributions are as follows: (i) we formalize a three-stage roadmap and distinguish deep research from related paradigms; (ii) we introduce four key components: query planning, information acquisition, memory management, and answer generation, each paired with fine-grained sub-taxonomies; (iii) we summarize optimization techniques, including prompting, supervised fine-tuning, and agentic reinforcement learning; and (iv) we consolidate evaluation criteria and open challenges, aiming to guide and facilitate future development. As the field of deep research continues to evolve rapidly, we are committed to continuously updating this survey to reflect the latest progress in this area.

Paperid: 630, https://arxiv.org/pdf/2512.01571.pdf

Abstract:
In this paper, we address the problem of fair access and Age of Information (AoI) optimization in 5G New Radio (NR) Vehicle to Everything (V2X) Mode 2. Specifically, vehicles need to exchange information with the road side unit (RSU). However, due to the varying vehicle speeds leading to different communication durations, the amount of data exchanged between different vehicles and the RSU may vary. This may poses significant safety risks in high-speed environments. To address this, we define a fairness index through tuning the selection window of different vehicles and consider the image semantic communication system to reduce latency. However, adjusting the selection window may affect the communication time, thereby impacting the AoI. Moreover, considering the re-evaluation mechanism in 5G NR, which helps reduce resource collisions, it may lead to an increase in AoI. We analyze the AoI using Stochastic Hybrid System (SHS) and construct a multi-objective optimization problem to achieve fair access and AoI optimization. Sequential Convex Approximation (SCA) is employed to transform the non-convex problem into a convex one, and solve it using convex optimization. We also provide a large language model (LLM) based algorithm. The scheme's effectiveness is validated through numerical simulations.

Paperid: 631, https://arxiv.org/pdf/2512.00805.pdf

Abstract:
Long video understanding is essential for human-like intelligence, enabling coherent perception and reasoning over extended temporal contexts. While the emerging thinking-with-frames paradigm, which alternates between global temporal reasoning and local frame examination, has advanced the reasoning capabilities of video multi-modal large language models (MLLMs), it suffers from a significant efficiency bottleneck due to the progressively growing and redundant multi-modal context. To address this, we propose SpecTemp, a reinforcement learning-based Speculative Temporal reasoning framework that decouples temporal perception from reasoning via a cooperative dual-model design. In SpecTemp, a lightweight draft MLLM rapidly explores and proposes salient frames from densely sampled temporal regions, while a powerful target MLLM focuses on temporal reasoning and verifies the draft's proposals, iteratively refining its attention until convergence. This design mirrors the collaborative pathways of the human brain, balancing efficiency with accuracy. To support training, we construct the SpecTemp-80K dataset, featuring synchronized dual-level annotations for coarse evidence spans and fine-grained frame-level evidence. Experiments across multiple video understanding benchmarks demonstrate that SpecTemp not only maintains competitive accuracy but also significantly accelerates inference compared with existing thinking-with-frames methods.

Paperid: 632, https://arxiv.org/pdf/2512.07522.pdf

Abstract:
Pre-training decoder-only language models relies on vast amounts of high-quality data, yet the availability of such data is increasingly reaching its limits. While metadata is commonly used to create and curate these datasets, its potential as a direct training signal remains under-explored. We challenge this status quo and propose LIME (Linguistic Metadata Embeddings), a method that enriches token embeddings with metadata capturing syntax, semantics, and contextual properties. LIME substantially improves pre-training efficiency. Specifically, it adapts up to 56% faster to the training data distribution, while introducing only 0.01% additional parameters at negligible compute overhead. Beyond efficiency, LIME improves tokenization, leading to remarkably stronger language modeling capabilities and generative task performance. These benefits persist across model scales (500M to 2B). In addition, we develop a variant with shifted metadata, LIME+1, that can guide token generation. Given prior metadata for the next token, LIME+1 improves reasoning performance by up to 38% and arithmetic accuracy by up to 35%.

Paperid: 633, https://arxiv.org/pdf/2512.05325.pdf

Abstract:
Large reasoning models achieve strong performance on complex tasks by generating extended chains of thought, but they often "overthink": continuing to reason long after they have enough information to answer correctly. This wastes inference-time compute and can hurt accuracy. Existing attempts to stop early either manipulate decoding with extra sampling and heuristics, rely on auxiliary verifier models, or operate only as post-hoc analysis pipelines without formal guarantees. We introduce LYNX, an online early-exit mechanism that turns a model's own hidden-state awareness into confidence-controlled stopping decisions. LYNX attaches exit decisions to naturally occurring reasoning cues (e.g., "hmm", "wait") during generation, trains a lightweight probe on hidden states at those cue tokens using supervision from forced exits, and wraps the resulting scores in split conformal prediction to obtain distribution-free control over premature exits. Crucially, we train and calibrate this probe once on a generic mathematical corpus and reuse it unchanged across benchmarks, decoding temperatures, and even non-mathematical tasks. Across three model families spanning 1.5B to 32B parameters, a single mathematically trained probe per base model yields strong accuracy--efficiency tradeoffs. On GSM8K, LYNX matches or improves baseline accuracy while reducing tokens by 40--65\%; on MATH-500 it improves accuracy by up to 12 points with roughly 35--60\% fewer tokens; on AIME 2024 it recovers baseline accuracy with more than 50\% token savings; and on CommonsenseQA, a non-math benchmark, it transfers zero-shot with modest accuracy gains and up to 70\% fewer tokens. Compared to state-of-the-art early-exit methods, LYNX offers competitive or superior Pareto frontiers while remaining fully online, requiring no proxy models at inference, and providing explicit, user-tunable confidence guarantees.

Paperid: 634, https://arxiv.org/pdf/2512.07650.pdf

Abstract:
Inspired by the success of language models (LM), scaling up deep learning recommendation systems (DLRS) has become a recent trend in the community. All previous methods tend to scale up the model parameters during training time. However, how to efficiently utilize and scale up computational resources during test time remains underexplored, which can prove to be a scaling-efficient approach and bring orthogonal improvements in LM domains. The key point in applying test-time scaling to DLRS lies in effectively generating diverse yet meaningful outputs for the same instance. We propose two ways: One is to explore the heterogeneity of different model architectures. The other is to utilize the randomness of model initialization under a homogeneous architecture. The evaluation is conducted across eight models, including both classic and SOTA models, on three benchmarks. Sufficient evidence proves the effectiveness of both solutions. We further prove that under the same inference budget, test-time scaling can outperform parameter scaling. Our test-time scaling can also be seamlessly accelerated with the increase in parallel servers when deployed online, without affecting the inference time on the user side. Code is available.

Paperid: 635, https://arxiv.org/pdf/2512.06868.pdf

Abstract:
Reliable incremental estimation of camera poses and 3D reconstruction is key to enable various applications including robotics, interactive visualization, and augmented reality. However, this task is particularly challenging in dynamic natural environments, where scene dynamics can severely deteriorate camera pose estimation accuracy. In this work, we propose a novel monocular visual SLAM system that can robustly estimate camera poses in dynamic scenes. To this end, we leverage the complementary strengths of geometric patch-based online bundle adjustment and recent feed-forward reconstruction models. Specifically, we propose a feed-forward reconstruction model to precisely filter out dynamic regions, while also utilizing its depth prediction to enhance the robustness of the patch-based visual SLAM. By aligning depth prediction with estimated patches from bundle adjustment, we robustly handle the inherent scale ambiguities of the batch-wise application of the feed-forward reconstruction model.

Paperid: 636, https://arxiv.org/pdf/2512.02834.pdf

Abstract:
Vision-Language-Action (VLA) models, trained via flow-matching or diffusion objectives, excel at learning complex behaviors from large-scale, multi-modal datasets (e.g., human teleoperation, scripted policies). However, since VLAs incorporate diverse data modes in the pre-training stage, and the finetuning dataset often contains demonstration data collected in a kinematically suboptimal or undesirable way, it exists redundant action modes that are irrelevant to the success action modes of the downstream task. Specifically, we observe a critical inference-time fragility among various sampled noises after supervised finetuning of pre-trained VLAs. In this paper, we attribute this instability to the distribution shift between the VLA policy and the policy induced by stable success modes of the downstream task dataset. Thus, we propose \textbf{TACO}, a test-time-scaling (TTS) framework that applies a lightweight pseudo-count estimator as a high-fidelity verifier of action chunks. The VLA models integrated with TACO can execute the actions with maximum pseudo-count from all sampled action chunks, thereby preventing distribution shifts while preserving the generalization ability of VLAs since the constraint is applied only during inference. Our method resembles the classical anti-exploration principle in offline reinforcement learning (RL), and being gradient-free, it incurs significant computational benefits compared to RL update, especially for flow or diffusion-based VLAs which are difficult to perform RL update due to denoising process. Extensive experiments across four simulation benchmarks (RoboTwin2.0, Robotwin, LIBERO, SimplerEnv) and a dual-arm platform demonstrate that our method significantly improves the inference stability and success rates in downstream-task adaptations.

Paperid: 637, https://arxiv.org/pdf/2512.03872.pdf

Abstract:
This paper investigates wireless systems aided by dual-polarized intelligent surfaces. We compare reconfigurable intelligent surface (RIS), which adjust their reflection matrices, with movable signals operating with fixed intelligent surface (FIS), which adjust the signal frequency while the surface properties remain fixed. For both RIS and FIS, we consider surfaces with a diagonal reflection matrix, named diagonal RIS/FIS, and surfaces with a reflection matrix not limited to being diagonal, named beyond-diagonal RIS/FIS. Movable signals with FIS always outperform RIS, achieving at least a fourfold gain. When transmitter and receiver polarizations differ, beyond-diagonal FIS further enhances performance.

Paperid: 638, https://arxiv.org/pdf/2512.03537.pdf

Abstract:
Existing class-incremental learning (CIL) approaches based on replay or knowledge distillation are often constrained by forgetting or the stability-plasticity dilemma. Some expansion-based approaches could achieve higher accuracy. However, they always require significant parameter increases. In this paper, we propose a plugin extension paradigm termed the Deployment of extra LoRA Components (DLC) for non-pre-trained CIL scenarios.We treat the feature extractor trained through replay or distillation as a base model with rich knowledge. For each task, we use Low-Rank Adaptation (LoRA) to inject task-specific residuals into the base model's deep layers. During inference, representations with task-specific residuals are aggregated to produce classification predictions. To mitigate interference from non-target LoRA plugins, we introduce a lightweight weighting unit. This unit learns to assign importance scores to different LoRA-tuned representations. Like downloadable contents in software, our method serves as a plug-and-play enhancement that efficiently extends the base methods. Remarkably, on the large-scale ImageNet-100, with merely 4 % of the parameters of a standard ResNet-18, our DLC model achieves a significant 8 % improvement in accuracy, demonstrating exceptional efficiency. Moreover, it could surpass state-of-the-art methods under the fixed memory budget.

Paperid: 639, https://arxiv.org/pdf/2512.01750.pdf

Abstract:
Integrated sensing and communication (ISAC) is a key enabler for low-altitude wireless networks (LAWNs), providing simultaneous environmental perception and data transmission in complex aerial scenarios. By combining heterogeneous sensing modalities such as visual, radar, lidar, and positional information, multimodal ISAC can improve both situational awareness and robustness of LAWNs. However, most existing multimodal fusion approaches use static fusion strategies that treat all modalities equally and cannot adapt to channel heterogeneity or time-varying modality reliability in dynamic low-altitude environments. To address this fundamental limitation, we propose a mixture-of-experts (MoE) framework for multimodal ISAC in LAWNs. Each modality is processed by a dedicated expert network, and a lightweight gating module adaptively assigns fusion weights according to the instantaneous informativeness and reliability of each modality. To improve scalability under the stringent energy constraints of aerial platforms, we further develop a sparse MoE variant that selectively activates only a subset of experts, thereby reducing computation overhead while preserving the benefits of adaptive fusion. Comprehensive simulations on three typical ISAC tasks in LAWNs demonstrate that the proposed frameworks consistently outperform conventional multimodal fusion baselines in terms of learning performance and training sample efficiency.

Paperid: 640, https://arxiv.org/pdf/2512.01427.pdf

Abstract:
Open-world and anomaly segmentation methods seek to enable autonomous driving systems to detect and segment both known and unknown objects in real-world scenes. However, existing methods do not assign semantically meaningful labels to unknown regions, and distinguishing and learning representations for unknown classes remains difficult. While open-vocabulary segmentation methods show promise in generalizing to novel classes, they require a fixed inference vocabulary and thus cannot be directly applied to anomaly segmentation where unknown classes are unconstrained. We propose Clipomaly, the first CLIP-based open-world and anomaly segmentation method for autonomous driving. Our zero-shot approach requires no anomaly-specific training data and leverages CLIP's shared image-text embedding space to both segment unknown objects and assign human-interpretable names to them. Unlike open-vocabulary methods, our model dynamically extends its vocabulary at inference time without retraining, enabling robust detection and naming of anomalies beyond common class definitions such as those in Cityscapes. Clipomaly achieves state-of-the-art performance on established anomaly segmentation benchmarks while providing interpretability and flexibility essential for practical deployment.

Paperid: 641, https://arxiv.org/pdf/2512.01210.pdf

Abstract:
Electronic health records (EHRs) support powerful clinical prediction models, but existing methods typically provide coarse, post hoc explanations that offer limited value for patient-level decision making. We introduce a knowledge graph (KG)-guided chain-of-thought (CoT) framework that generates clinically grounded and temporally consistent reasoning for visit-level disease prediction in MIMIC-III. ICD-9 codes are mapped to PrimeKG, from which disease-relevant nodes and multi-hop reasoning paths are extracted and used as scaffolds for CoT generation; only explanations whose conclusions match observed outcomes are retained. Lightweight LLaMA-3.1-Instruct-8B and Gemma-7B models are then fine-tuned on this supervision corpus. Across ten PrimeKG-mapped diseases and limited training cohorts (400 and 1000 cases), KG-guided models outperform strong classical baselines, achieving AUROC values of 0.66 to 0.70 and macro-AUPR values of 0.40 to 0.47. The models also transfer zero-shot to the CRADLE cohort, improving accuracy from approximately 0.40 to 0.51 up to 0.72 to 0.77. A blinded clinician evaluation shows consistent preference for KG-guided CoT explanations in clarity, relevance, and clinical correctness.

Paperid: 642, https://arxiv.org/pdf/2512.06151.pdf

Abstract:
This paper presents a real-time control framework for nonlinear pure-feedback systems with unknown dynamics to satisfy reach-avoid-stay tasks within a prescribed time in dynamic environments. To achieve this, we introduce a real-time spatiotemporal tube (STT) framework. An STT is defined as a time-varying ball in the state space whose center and radius adapt online using only real-time sensory input. A closed-form, approximation-free control law is then derived to constrain the system output within the STT, ensuring safety and task satisfaction. We provide formal guarantees for obstacle avoidance and on-time task completion. The effectiveness and scalability of the framework are demonstrated through simulations and hardware experiments on a mobile robot and an aerial vehicle, navigating in cluttered dynamic environments.

Paperid: 643, https://arxiv.org/pdf/2512.05495.pdf

Abstract:
This paper presents a Spatiotemporal Tube (STT)-based control framework for differential-drive mobile robots with dynamic uncertainties and external disturbances, guaranteeing the satisfaction of Temporal Reach-Avoid-Stay (T-RAS) specifications. The approach employs circular STT, characterized by smoothly time-varying center and radius, to define dynamic safe corridors that guide the robot from the start region to the goal while avoiding obstacles. In particular, we first develop a sampling-based synthesis algorithm to construct a feasible STT that satisfies the prescribed timing and safety constraints with formal guarantees. To ensure that the robot remains confined within this tube, we then design analytically a closed-form, approximation-free control law. The resulting controller is computationally efficient, robust to disturbances and {model uncertainties}, and requires no model approximations or online optimization. The proposed framework is validated through simulation studies on a differential-drive robot and benchmarked against state-of-the-art methods, demonstrating superior robustness, accuracy, and computational efficiency.

Paperid: 644, https://arxiv.org/pdf/2512.04949.pdf

Abstract:
Agents capable of accomplishing complex tasks through multiple interactions with the environment have emerged as a popular research direction. However, in such multi-step settings, the conventional group-level policy optimization algorithm becomes suboptimal because of its underlying assumption that each action holds equal contribution, which deviates significantly from reality. Our analysis reveals that only a small fraction of actions are critical in determining the final outcome. Building on this insight, we propose CARL, a critical-action-focused reinforcement learning algorithm tailored for multi-step agents. CARL achieves focused training through providing action-level optimization signals for high-criticality actions while excluding low-criticality actions from model update. Extensive experiments demonstrate that CARL achieves both stronger performance and higher efficiency during training and inference across diverse evaluation settings.

Paperid: 645, https://arxiv.org/pdf/2512.03197.pdf

Abstract:
Large Language Models (LLMs) have revolutionized the ability to understand and generate text, enabling significant progress in automatic knowledge graph construction from text (Text2KG). Many Text2KG methods, however, rely on iterative LLM prompting, making them computationally expensive and prone to overlooking complex relations distributed throughout the text. To address these limitations, we propose InvertiTune, a framework that combines a controlled data generation pipeline with supervised fine-tuning (SFT). Within this framework, the data-generation pipeline systematically extracts subgraphs from large knowledge bases, applies noise filtering, and leverages LLMs to generate corresponding natural text descriptions, a task more aligned with LLM capabilities than direct KG generation from text. This pipeline enables generating datasets composed of longer texts paired with larger KGs that better reflect real-world scenarios compared to existing benchmarks, thus supporting effective SFT of lightweight models for single-shot KG construction. Experimental results on CE12k, a dataset generated using the introduced pipeline, show that InvertiTune outperforms larger non-fine-tuned LLMs as well as state-of-the-art Text2KG approaches, while also demonstrating stronger cross-dataset generalization on CrossEval-1200, a test set created from three established benchmark datasets and CE12k. These findings highlight the importance of realistic, high-quality training data for advancing efficient and high-performing Text2KG systems.

Paperid: 646, https://arxiv.org/pdf/2512.02460.pdf

Abstract:
Searching and detecting communities in real-world graphs underpins a wide range of applications. Despite the success achieved, current learning-based solutions regard community search, i.e., locating the best community for a given query, and community detection, i.e., partitioning the whole graph, as separate problems, necessitating task- and dataset-specific retraining. Such a strategy limits the applicability and generalization ability of the existing models. Additionally, these methods rely heavily on information from the target dataset, leading to suboptimal performance when supervision is limited or unavailable. To mitigate this limitation, we propose UniCom, a unified framework to solve both community search and detection tasks through knowledge transfer across multiple domains, thus alleviating the limitations of single-dataset learning. UniCom centers on a Domain-aware Specialization (DAS) procedure that adapts on the fly to unseen graphs or tasks, eliminating costly retraining while maintaining framework compactness with a lightweight prompt-based paradigm. This is empowered by a Universal Graph Learning (UGL) backbone, which distills transferable semantic and topological knowledge from multiple source domains via comprehensive pre-training. Both DAS and UGL are informed by local neighborhood signals and cohesive subgraph structures, providing consistent guidance throughout the framework. Extensive experiments on both tasks across 16 benchmark datasets and 22 baselines have been conducted to ensure a comprehensive and fair evaluation. UniCom consistently outperforms all state-of-the-art baselines across all tasks under settings with scarce or no supervision, while maintaining runtime efficiency.

Paperid: 647, https://arxiv.org/pdf/2512.06769.pdf

Abstract:
Existing vision-language models often suffer from spatial hallucinations, i.e., generating incorrect descriptions about the relative positions of objects in an image. We argue that this problem mainly stems from the asymmetric properties between images and text. To enrich the spatial understanding ability of vision-language models, we propose a simple, annotation-free, plug-and-play method named $\text{Stitch and Tell}$ (abbreviated as SiTe), which injects structured spatial supervision into data. It constructs stitched image-text pairs by stitching images along a spatial axis and generating spatially-aware captions or question answer pairs based on the layout of stitched image, without relying on costly advanced models or human involvement. We evaluate SiTe across three architectures including LLaVA-v1.5-7B, LLaVA-Qwen2-1.5B and HALVA-7B, two training datasets, and eight benchmarks. Experiments show that SiTe improves spatial understanding tasks such as $\text{MME}_{\text{Position}}$ (+5.50%) and Spatial-MM (+4.19%), while maintaining or improving performance on general vision-language benchmarks including COCO-QA (+1.02%) and MMBench (+4.76%). Our findings suggest that explicitly injecting spatially-aware structure into training data offers an effective way to mitigate spatial hallucinations and improve spatial understanding, while preserving general vision-language capabilities.

Paperid: 648, https://arxiv.org/pdf/2512.06381.pdf

Abstract:
The parallelized multi-retrieval architecture has been widely adopted in large-scale recommender systems for its computational efficiency and comprehensive coverage of user interests. Many retrieval methods typically integrate additional cross-scenario samples to enhance the overall performance ceiling. However, those model designs neglect the fact that a part of the cross-scenario samples have already been retrieved by existing models within a system, leading to diminishing marginal utility in delivering incremental performance gains. In this paper, we propose a novel retrieval framework IncRec, specifically for cross-scenario incremental sample learning. The innovations of IncRec can be highlighted as two aspects. Firstly, we construct extreme cross-scenario incremental samples that are not retrieved by any existing model. And we design an incremental sample learning framework which focuses on capturing incremental representation to improve the overall retrieval performance. Secondly, we introduce a consistency-aware alignment module to further make the model prefer incremental samples with high exposure probability. Extensive offline and online A/B tests validate the superiority of our framework over state-of-the-art retrieval methods. In particular, we deploy IncRec in the Taobao homepage recommendation, achieving a 1% increase in online transaction count, demonstrating its practical applicability.

Paperid: 649, https://arxiv.org/pdf/2512.02321.pdf

Abstract:
Large Language Model (LLM)-based agents have demonstrated remarkable capabilities in reasoning, planning, and tool usage. The recently proposed Model Context Protocol (MCP) has emerged as a unifying framework for integrating external tools into agent systems, enabling a thriving open ecosystem of community-built functionalities. However, the openness and composability that make MCP appealing also introduce a critical yet overlooked security assumption -- implicit trust in third-party tool providers. In this work, we identify and formalize a new class of attacks that exploit this trust boundary without violating explicit permissions. We term this new attack vector implicit toxicity, where malicious behaviors occur entirely within the allowed privilege scope. We propose LeechHijack, a Latent Embedded Exploit for Computation Hijacking, in which an adversarial MCP tool covertly expropriates the agent's computational resources for unauthorized workloads. LeechHijack operates through a two-stage mechanism: an implantation stage that embeds a benign-looking backdoor in a tool, and an exploitation stage where the backdoor activates upon predefined triggers to establish a command-and-control channel. Through this channel, the attacker injects additional tasks that the agent executes as if they were part of its normal workflow, effectively parasitizing the user's compute budget. We implement LeechHijack across four major LLM families. Experiments show that LeechHijack achieves an average success rate of 77.25%, with a resource overhead of 18.62% compared to the baseline. This study highlights the urgent need for computational provenance and resource attestation mechanisms to safeguard the emerging MCP ecosystem.

Paperid: 650, https://arxiv.org/pdf/2512.01300.pdf

Abstract:
Current Vision-Language Model (VLM)-based end-to-end autonomous driving systems often leverage large language models to generate driving decisions directly based on their understanding of the current scene. However, such systems introduce multiple risks in real-world driving scenarios. To evaluate whether VLMs are truly viable for autonomous driving, we introduce RoboDriveBench, the first robustness benchmark focused on end-to-end trajectory prediction tasks. This benchmark systematically evaluates two critical categories of real-world challenges for VLM-based end-to-end autonomous driving systems through 11 simulated scenarios encompassing various corruption types, including 6 scenarios of sensor corruption caused by environmental variations, along with 5 cases of prompt corruption resulting from human intervention and data transmission failures. Each corruption type includes 250 unique driving scenarios and 5,689 frames, resulting in 64,559 total trajectory prediction cases per evaluation. To overcome these real-world challenges, we propose a novel VLM-based autonomous driving framework called RoboDriveVLM, which enhances robustness by mapping more multimodal data-e.g., lidar and radar-into a unified latent space. Furthermore, we introduce a new Test-Time Adaptation (TTA) method based on cross-modal knowledge distillation to improve the robustness of VLM-based autonomous driving systems. Through extensive experiments, our work highlights the limitations of current VLM-based end-to-end autonomous driving systems and provides a more reliable solution for real-world deployment. Source code and datasets will be released.

Paperid: 651, https://arxiv.org/pdf/2512.01273.pdf

Abstract:
Retinal imaging is a critical, non-invasive modality for the early detection and monitoring of ocular and systemic diseases. Deep learning, particularly convolutional neural networks (CNNs), has significant progress in automated retinal analysis, supporting tasks such as fundus image classification, lesion detection, and vessel segmentation. As a representative lightweight network, nnMobileNet has demonstrated strong performance across multiple retinal benchmarks while remaining computationally efficient. However, purely convolutional architectures inherently struggle to capture long-range dependencies and model the irregular lesions and elongated vascular patterns that characterize on retinal images, despite the critical importance of vascular features for reliable clinical diagnosis. To further advance this line of work and extend the original vision of nnMobileNet, we propose nnMobileNet++, a hybrid architecture that progressively bridges convolutional and transformer representations. The framework integrates three key components: (i) dynamic snake convolution for boundary-aware feature extraction, (ii) stage-specific transformer blocks introduced after the second down-sampling stage for global context modeling, and (iii) retinal image pretraining to improve generalization. Experiments on multiple public retinal datasets for classification, together with ablation studies, demonstrate that nnMobileNet++ achieves state-of-the-art or highly competitive accuracy while maintaining low computational cost, underscoring its potential as a lightweight yet effective framework for retinal image analysis.

Paperid: 652, https://arxiv.org/pdf/2512.07582.pdf

Abstract:
Developing robust and general-purpose manipulation policies represents a fundamental objective in robotics research. While Vision-Language-Action (VLA) models have demonstrated promising capabilities for end-to-end robot control, existing approaches still exhibit limited generalization to tasks beyond their training distributions. In contrast, humans possess remarkable proficiency in acquiring novel skills by simply observing others performing them once. Inspired by this capability, we propose ViVLA, a generalist robotic manipulation policy that achieves efficient task learning from a single expert demonstration video at test time. Our approach jointly processes an expert demonstration video alongside the robot's visual observations to predict both the demonstrated action sequences and subsequent robot actions, effectively distilling fine-grained manipulation knowledge from expert behavior and transferring it seamlessly to the agent. To enhance the performance of ViVLA, we develop a scalable expert-agent pair data generation pipeline capable of synthesizing paired trajectories from easily accessible human videos, further augmented by curated pairs from publicly available datasets. This pipeline produces a total of 892,911 expert-agent samples for training ViVLA. Experimental results demonstrate that our ViVLA is able to acquire novel manipulation skills from only a single expert demonstration video at test time. Our approach achieves over 30% improvement on unseen LIBERO tasks and maintains above 35% gains with cross-embodiment videos. Real-world experiments demonstrate effective learning from human videos, yielding more than 38% improvement on unseen tasks.

Paperid: 653, https://arxiv.org/pdf/2512.05126.pdf

Abstract:
Video dubbing aims to generate high-fidelity speech that is precisely temporally aligned with the visual content. Existing methods still suffer from limitations in speech naturalness and audio-visual synchronization, and are limited to monolingual settings. To address these challenges, we propose SyncVoice, a vision-augmented video dubbing framework built upon a pretrained text-to-speech (TTS) model. By fine-tuning the TTS model on audio-visual data, we achieve strong audiovisual consistency. We propose a Dual Speaker Encoder to effectively mitigate inter-language interference in cross-lingual speech synthesis and explore the application of video dubbing in video translation scenarios. Experimental results show that SyncVoice achieves high-fidelity speech generation with strong synchronization performance, demonstrating its potential in video dubbing tasks.

Paperid: 654, https://arxiv.org/pdf/2512.04529.pdf

Abstract:
Generating academic slides from scientific papers is a challenging multimodal reasoning task that requires both long context understanding and deliberate visual planning. Existing approaches largely reduce it to text only summarization, overlooking the visual component and design intensive nature of slide creation. In this paper we introduce SlideGen, an agentic, modular, and visual in the loop framework for scientific paper to slide generation. SlideGen orchestrates a group of vision language agents that reason collaboratively over the document structure and semantics, producing editable PPTX slides with logical flow and compelling visual presentation. By integrating coordinated outlining, mapping, arrangement, note synthesis, and iterative refinement, our system consistently delivers slides of expert level quality. Across diverse benchmarks and strong baselines, SlideGen outperforms existing methods in visual quality, content faithfulness, and readability, positioning it as the new state of the art in automated slide generation. Our work establishes a foundation for design aware multimodal slide generation, demonstrating how agentic collaboration can bridge understanding and presentation in complex multimodal reasoning tasks.

Paperid: 655, https://arxiv.org/pdf/2512.07480.pdf

Abstract:
While traditional and neural video codecs (NVCs) have achieved remarkable rate-distortion performance, improving perceptual quality at low bitrates remains challenging. Some NVCs incorporate perceptual or adversarial objectives but still suffer from artifacts due to limited generation capacity, whereas others leverage pretrained diffusion models to improve quality at the cost of heavy sampling complexity. To overcome these challenges, we propose S2VC, a Single-Step diffusion based Video Codec that integrates a conditional coding framework with an efficient single-step diffusion generator, enabling realistic reconstruction at low bitrates with reduced sampling cost. Recognizing the importance of semantic conditioning in single-step diffusion, we introduce Contextual Semantic Guidance to extract frame-adaptive semantics from buffered features. It replaces text captions with efficient, fine-grained conditioning, thereby improving generation realism. In addition, Temporal Consistency Guidance is incorporated into the diffusion U-Net to enforce temporal coherence across frames and ensure stable generation. Extensive experiments show that S2VC delivers state-of-the-art perceptual quality with an average 52.73% bitrate saving over prior perceptual methods, underscoring the promise of single-step diffusion for efficient, high-quality video compression.

Paperid: 656, https://arxiv.org/pdf/2512.05905.pdf

Abstract:
Achieving character animation that meets studio-grade production standards remains challenging despite recent progress. Existing approaches can transfer motion from a driving video to a reference image, but often fail to preserve structural fidelity and temporal consistency in wild scenarios involving complex motion and cross-identity animations. In this work, we present \textbf{SCAIL} (\textbf{S}tudio-grade \textbf{C}haracter \textbf{A}nimation via \textbf{I}n-context \textbf{L}earning), a framework designed to address these challenges from two key innovations. First, we propose a novel 3D pose representation, providing a more robust and flexible motion signal. Second, we introduce a full-context pose injection mechanism within a diffusion-transformer architecture, enabling effective spatio-temporal reasoning over full motion sequences. To align with studio-level requirements, we develop a curated data pipeline ensuring both diversity and quality, and establish a comprehensive benchmark for systematic evaluation. Experiments show that \textbf{SCAIL} achieves state-of-the-art performance and advances character animation toward studio-grade reliability and realism.

Paperid: 657, https://arxiv.org/pdf/2512.04668.pdf

Abstract:
Graph topology is a fundamental determinant of memory leakage in multi-agent LLM systems, yet its effects remain poorly quantified. We introduce MAMA (Multi-Agent Memory Attack), a framework that measures how network structure shapes leakage. MAMA operates on synthetic documents containing labeled Personally Identifiable Information (PII) entities, from which we generate sanitized task instructions. We execute a two-phase protocol: Engram (seeding private information into a target agent's memory) and Resonance (multi-round interaction where an attacker attempts extraction). Over up to 10 interaction rounds, we quantify leakage as the fraction of ground-truth PII recovered from attacking agent outputs via exact matching. We systematically evaluate six common network topologies (fully connected, ring, chain, binary tree, star, and star-ring), varying agent counts $n\in\{4,5,6\}$, attacker-target placements, and base models. Our findings reveal consistent patterns: fully connected graphs exhibit maximum leakage while chains provide strongest protection; shorter attacker-target graph distance and higher target centrality significantly increase vulnerability; leakage rises sharply in early rounds before plateauing; model choice shifts absolute leakage rates but preserves topology rankings; temporal/locational PII attributes leak more readily than identity credentials or regulated identifiers. These results provide the first systematic mapping from architectural choices to measurable privacy risk, yielding actionable guidance: prefer sparse or hierarchical connectivity, maximize attacker-target separation, limit node degree and network radius, avoid shortcuts bypassing hubs, and implement topology-aware access controls.

Paperid: 658, https://arxiv.org/pdf/2512.04538.pdf

Abstract:
As code completion task from function-level to repository-level, leveraging contextual information from large-scale codebases becomes a core challenge. However, existing retrieval-augmented generation (RAG) methods typically treat code as plain natural language, relying primarily on shallow semantic matching while overlooking structural semantics and code-specific dependencies. This limits their ability to capture control flow and underlying intent, ultimately constraining the quality of generated code. Therefore, we propose CoCo, a novel framework that enables code Completion by Comprehension of multi-granularity context from large-scale code repositories. CoCo employs static code analysis to extract structured context at the function, file, and project levels, capturing execution logic and semantic dependencies. It then adopts an graph-based multi-granularity context selection mechanism to filter out redundant information and remove noise. Consequently, the information is converted into natural language in a consistent manner, thereby functioning as explicit contextual prompts to guide subsequent code completion. Additionally, a structure-aware code re-ranker mechanism ensures alignment at both semantic and structural levels. Extensive experiments on CrossCodeEval and RepoEval benchmarks demonstrate that CoCo consistently surpasses state-of-the-art baselines, achieving up to 20.2% gains in EM. Moreover, the framework is model-agnostic and can be seamlessly integrated into existing methods, leading to significant performance.

Paperid: 659, https://arxiv.org/pdf/2512.04000.pdf

Abstract:
The application of Large Multimodal Models (LMMs) to long-form video understanding is constrained by limited context lengths and the computationally prohibitive cost of processing dense video tokens. Consequently, recent research has focused on query-aware frame selection, methods that often incur significant computational overhead. This paper challenges the assumption that such complex search mechanisms are universally necessary. We first identify and validate a query typology distinguishing between global query and localized query. We demonstrate that while uniform sampling is both effective and efficient for global queries, localized queries indeed necessitate query-aware selection for optimal performance. Building on this insight, we propose DIG, a training-free frame selection framework that adapts its strategy based on the query type. Specifically,DIG employs efficient uniform sampling for global queries while activating a specialized pipeline to extract query-relevant frames for localized queries. Experiments on three long-form video understanding benchmarks demonstrate that DIG consistently outperforms existing baselines and robustly improves LMM performance, even when scaling the input frame count to 256.

Paperid: 660, https://arxiv.org/pdf/2512.03593.pdf

Abstract:
We present a CloseUpAvatar - a novel approach for articulated human avatar representation dealing with more general camera motions, while preserving rendering quality for close-up views. CloseUpAvatar represents an avatar as a set of textured planes with two sets of learnable textures for low and high-frequency detail. The method automatically switches to high-frequency textures only for cameras positioned close to the avatar's surface and gradually reduces their impact as the camera moves farther away. Such parametrization of the avatar enables CloseUpAvatar to adjust rendering quality based on camera distance ensuring realistic rendering across a wider range of camera orientations than previous approaches. We provide experiments using the ActorsHQ dataset with high-resolution input images. CloseUpAvatar demonstrates both qualitative and quantitative improvements over existing methods in rendering from novel wide range camera positions, while maintaining high FPS by limiting the number of required primitives.

Paperid: 661, https://arxiv.org/pdf/2512.02814.pdf

Abstract:
Radiology reporting is an essential yet time-consuming and error-prone task for radiologists in clinical examinations, especially for volumetric medical images. Rigorous quality control is also critical but tedious, ensuring that the final report meets clinical standards. Existing automated approaches, including radiology report generation methods and medical vision-language models, focus mainly on the report generation phase and neglect the crucial quality control procedure, limiting their capability to provide comprehensive support to radiologists. We propose Radiologist Copilot, an agentic AI assistant equipped with orchestrated tools designed for automated radiology reporting with quality control. Leveraging large language models as the reasoning backbone, the agentic system autonomously selects tools, plans, and executes actions, emulating the behavior of radiologists throughout the holistic radiology reporting process. The orchestrated tools include region localization, think with image paradigm directed region analysis planning, strategic template selection for report generation, quality assessment and feedback-driven adaptive refinement for quality control. Therefore, Radiologist Copilot facilitates accurate, complete, and efficient radiology reporting, assisting radiologists and improving clinical efficiency. Experimental results demonstrate that Radiologist Copilot significantly surpasses other state-of-the-art methods in radiology reporting. The source code will be released upon acceptance.

Paperid: 662, https://arxiv.org/pdf/2512.02423.pdf

Abstract:
With the rapid development of Large Vision Language Models, the focus of Graphical User Interface (GUI) agent tasks shifts from single-screen tasks to complex screen navigation challenges. However, real-world GUI environments, such as PC software and mobile Apps, are often complex and proprietary, making it difficult to obtain the comprehensive environment information needed for agent training and evaluation. This limitation hinders systematic investigation and benchmarking of agent navigation capabilities. To address this limitation, we introduce GUI Exploration Lab, a simulation environment engine for GUI agent navigation research that enables flexible definition and composition of screens, icons, and navigation graphs, while providing full access to environment information for comprehensive agent training and evaluation. Through extensive experiments, we find that supervised fine-tuning enables effective memorization of fundamental knowledge, serving as a crucial foundation for subsequent training. Building on this, single-turn reinforcement learning further enhances generalization to unseen scenarios. Finally, multi-turn reinforcement learning encourages the development of exploration strategies through interactive trial and error, leading to further improvements in screen navigation performance. We validate our methods on both static and interactive benchmarks, demonstrating that our findings generalize effectively to real-world scenarios. These findings demonstrate the advantages of reinforcement learning approaches in GUI navigation and offer practical guidance for building more capable and generalizable GUI agents.

Paperid: 663, https://arxiv.org/pdf/2512.00873.pdf

Abstract:
Cone beam computed tomography (CBCT)-guided puncture has become an established approach for diagnosing and treating early- to mid-stage thoracic tumours, yet the associated radiation exposure substantially elevates the risk of secondary malignancies. Although multiple low-dose CBCT strategies have been introduced, none have undergone validation using large-scale multicenter retrospective datasets, and prospective clinical evaluation remains lacking. Here, we propose DeepPriorCBCT - a three-stage deep learning framework that achieves diagnostic-grade reconstruction using only one-sixth of the conventional radiation dose. 4102 patients with 8675 CBCT scans from 12 centers were included to develop and validate DeepPriorCBCT. Additionally, a prospective cross-over trial (Registry number: NCT07035977) which recruited 138 patients scheduled for percutaneous thoracic puncture was conducted to assess the model's clinical applicability. Assessment by 11 physicians confirmed that reconstructed images were indistinguishable from original scans. Moreover, diagnostic performance and overall image quality were comparable to those generated by standard reconstruction algorithms. In the prospective trial, five radiologists reported no significant differences in image quality or lesion assessment between DeepPriorCBCT and the clinical standard (all P>0.05). Likewise, 25 interventionalists expressed no preference between model-based and full-sampling images for surgical guidance (Kappa<0.2). Radiation exposure with DeepPriorCBCT was reduced to approximately one-sixth of that with the conventional approach, and collectively, the findings confirm that it enables high-quality CBCT reconstruction under sparse sampling conditions while markedly decreasing intraoperative radiation risk.

Paperid: 664, https://arxiv.org/pdf/2512.00723.pdf

Abstract:
End-to-end autonomous driving systems directly generate driving policies from raw sensor inputs. While these systems can extract effective environmental features for planning, relying on auxiliary perception tasks, developing perception annotation-free planning paradigms has become increasingly critical due to the high cost of manual perception annotation. In this work, we propose TrajDiff, a Trajectory-oriented BEV Conditioned Diffusion framework that establishes a fully perception annotation-free generative method for end-to-end autonomous driving. TrajDiff requires only raw sensor inputs and future trajectory, constructing Gaussian BEV heatmap targets that inherently capture driving modalities. We design a simple yet effective trajectory-oriented BEV encoder to extract the TrajBEV feature without perceptual supervision. Furthermore, we introduce Trajectory-oriented BEV Diffusion Transformer (TB-DiT), which leverages ego-state information and the predicted TrajBEV features to directly generate diverse yet plausible trajectories, eliminating the need for handcrafted motion priors. Beyond architectural innovations, TrajDiff enables exploration of data scaling benefits in the annotation-free setting. Evaluated on the NAVSIM benchmark, TrajDiff achieves 87.5 PDMS, establishing state-of-the-art performance among all annotation-free methods. With data scaling, it further improves to 88.5 PDMS, which is comparable to advanced perception-based approaches. Our code and model will be made publicly available.

Paperid: 665, https://arxiv.org/pdf/2512.00400.pdf

Abstract:
The rapid evolution of edge computing has exposed fundamental limitations in traditional operating system and hypervisor architectures, particularly in managing heterogeneous platforms and meeting the constraints of limited resources. Existing solutions often rely on monolithic or layered combinations of hypervisors and guest OSes, which are difficult to tailor for the diverse and dynamic requirements of edge scenarios. To address these challenges, we propose TenonOS, a demand-driven, self-generating, and lightweight operating system framework that fundamentally rethinks and reconstructs both the hypervisor and OS architectures. TenonOS introduces a novel LibOS-on-LibOS approach, in which both virtualization and OS functionalities are modularized into fine-grained, reusable micro-libraries. A dynamic orchestration engine composes these modules on demand to construct customized, application-specific runtime environments. At the core of TenonOS are two key components: Mortise, a minimal, modularized hypervisor, and Tenon, a real-time LibOS. Mortise provides low-overhead resource isolation, fast inter-VM communication, and manages the full lifecycle of Tenon instances - including on-demand creation, suspension, and termination - enabling TenonOS to flexibly adapt its runtime layout to workload variations. Tenon delivers deterministic scheduling and multi-process support for time-critical applications. Through this unified and modular architecture, TenonOS eliminates redundant layers, reduces system overhead, and enhances scalability, security, and maintainability. Extensive evaluations demonstrate that TenonOS achieves superior real-time scheduling (40.28% improvement), a compact memory footprint (361 KiB), and high adaptability to dynamic edge workloads, making it an ideal foundation for heterogeneous, resource-constrained edge systems.

Paperid: 666, https://arxiv.org/pdf/2512.07507.pdf

Abstract:
The rapid development of autonomous vehicles has led to a surge in testing demand. Traditional testing methods, such as virtual simulation, closed-course, and public road testing, face several challenges, including unrealistic vehicle states, limited testing capabilities, and high costs. These issues have prompted increasing interest in virtual-physical fusion testing. However, despite its potential, virtual-physical fusion testing still faces challenges, such as limited element types, narrow testing scope, and fixed evaluation metrics. To address these challenges, we propose the Virtual-Physical Testing Platform for Autonomous Vehicles (VP-AutoTest), which integrates over ten types of virtual and physical elements, including vehicles, pedestrians, and roadside infrastructure, to replicate the diversity of real-world traffic participants. The platform also supports both single-vehicle interaction and multi-vehicle cooperation testing, employing adversarial testing and parallel deduction to accelerate fault detection and explore algorithmic limits, while OBU and Redis communication enable seamless vehicle-to-vehicle (V2V) and vehicle-to-infrastructure (V2I) cooperation across all levels of cooperative automation. Furthermore, VP-AutoTest incorporates a multidimensional evaluation framework and AI-driven expert systems to conduct comprehensive performance assessment and defect diagnosis. Finally, by comparing virtual-physical fusion test results with real-world experiments, the platform performs credibility self-evaluation to ensure both the fidelity and efficiency of autonomous driving testing. Please refer to the website for the full testing functionalities on the autonomous driving public service platform OnSite:https://www.onsite.com.cn.

Paperid: 667, https://arxiv.org/pdf/2512.07135.pdf

Abstract:
Current autonomous driving systems often favor end-to-end frameworks, which take sensor inputs like images and learn to map them into trajectory space via neural networks. Previous work has demonstrated that models can achieve better planning performance when provided with a prior distribution of possible trajectories. However, these approaches often overlook two critical aspects: 1) The appropriate trajectory prior can vary significantly across different driving scenarios. 2) Their trajectory evaluation mechanism lacks policy-driven refinement, remaining constrained by the limitations of one-stage supervised training. To address these issues, we explore improvements in two key areas. For problem 1, we employ MoE to apply different trajectory priors tailored to different scenarios. For problem 2, we utilize Reinforcement Learning to fine-tune the trajectory scoring mechanism. Additionally, we integrate models with different perception backbones to enhance perceptual features. Our integrated model achieved a score of 51.08 on the navsim ICCV benchmark, securing third place.

Paperid: 668, https://arxiv.org/pdf/2512.06106.pdf

Abstract:
What if users could meet their future selves today? AI-generated future selves simulate meaningful encounters with a digital twin decades in the future. As AI systems advance, combining cloned voices, age-progressed facial rendering, and autobiographical narratives, a central question emerges: Does the modality of these future selves alter their psychological and affective impact? How might a text-based chatbot, a voice-only system, or a photorealistic avatar shape present-day decisions and our feeling of connection to the future? We report a randomized controlled study (N=92) evaluating three modalities of AI-generated future selves (text, voice, avatar) against a neutral control condition. We also report a systematic model evaluation between Claude 4 and three other Large Language Models (LLMs), assessing Claude 4 across psychological and interaction dimensions and establishing conversational AI quality as a critical determinant of intervention effectiveness. All personalized modalities strengthened Future Self-Continuity (FSC), emotional well-being, and motivation compared to control, with avatar producing the largest vividness gains, yet with no significant differences between formats. Interaction quality metrics, particularly persuasiveness, realism, and user engagement, emerged as robust predictors of psychological and affective outcomes, indicating that how compelling the interaction feels matters more than the form it takes. Content analysis found thematic patterns: text emphasized career planning, while voice and avatar facilitated personal reflection. Claude 4 outperformed ChatGPT 3.5, Llama 4, and Qwen 3 in enhancing psychological, affective, and FSC outcomes.

Paperid: 669, https://arxiv.org/pdf/2512.05853.pdf

Abstract:
Multimodal Large Language Models (MLLMs) are widely used in various fields due to their powerful cross-modal comprehension and generation capabilities. However, more modalities bring more vulnerabilities to being utilized for jailbreak attacks, which induces MLLMs to output harmful content. Due to the strong reasoning ability of MLLMs, previous jailbreak attacks try to explore reasoning safety risk in text modal, while similar threats have been largely overlooked in the visual modal. To fully evaluate potential safety risks in the visual reasoning task, we propose Visual Reasoning Sequential Attack (VRSA), which induces MLLMs to gradually externalize and aggregate complete harmful intent by decomposing the original harmful text into several sequentially related sub-images. In particular, to enhance the rationality of the scene in the image sequence, we propose Adaptive Scene Refinement to optimize the scene most relevant to the original harmful query. To ensure the semantic continuity of the generated image, we propose Semantic Coherent Completion to iteratively rewrite each sub-text combined with contextual information in this scene. In addition, we propose Text-Image Consistency Alignment to keep the semantical consistency. A series of experiments demonstrates that the VRSA can achieve a higher attack success rate compared with the state-of-the-art jailbreak attack methods on both the open-source and closed-source MLLMs such as GPT-4o and Claude-4.5-Sonnet.

Paperid: 670, https://arxiv.org/pdf/2512.05033.pdf

Abstract:
Modern Large Language Models achieve impressive reasoning capabilities with long Chain of Thoughts, but they incur substantial computational cost during inference, and this motivates techniques to improve the performance-cost ratio. Among these techniques, Speculative Decoding accelerates inference by employing a fast but inaccurate draft model to autoregressively propose tokens, which are then verified in parallel by a more capable target model. However, due to unnecessary rejections caused by token mismatches in semantically equivalent steps, traditional token-level Speculative Decoding struggles in reasoning tasks. Although recent works have shifted to step-level semantic verification, which improve efficiency by accepting or rejecting entire reasoning steps, existing step-level methods still regenerate many rejected steps with little improvement, wasting valuable target compute. To address this challenge, we propose Arbitrage, a novel step-level speculative generation framework that routes generation dynamically based on the relative advantage between draft and target models. Instead of applying a fixed acceptance threshold, Arbitrage uses a lightweight router trained to predict when the target model is likely to produce a meaningfully better step. This routing approximates an ideal Arbitrage Oracle that always chooses the higher-quality step, achieving near-optimal efficiency-accuracy trade-offs. Across multiple mathematical reasoning benchmarks, Arbitrage consistently surpasses prior step-level Speculative Decoding baselines, reducing inference latency by up to $\sim2\times$ at matched accuracy.

Paperid: 671, https://arxiv.org/pdf/2512.04772.pdf

Abstract:
In recent years, precision agriculture has been introducing groundbreaking innovations in the field, with a strong focus on automation. However, research studies in robotics and autonomous navigation often rely on controlled simulations or isolated field trials. The absence of a realistic common benchmark represents a significant limitation for the diffusion of robust autonomous systems under real complex agricultural conditions. Vineyards pose significant challenges due to their dynamic nature, and they are increasingly drawing attention from both academic and industrial stakeholders interested in automation. In this context, we introduce the TEMPO-VINE dataset, a large-scale multi-temporal dataset specifically designed for evaluating sensor fusion, simultaneous localization and mapping (SLAM), and place recognition techniques within operational vineyard environments. TEMPO-VINE is the first multi-modal public dataset that brings together data from heterogeneous LiDARs of different price levels, AHRS, RTK-GPS, and cameras in real trellis and pergola vineyards, with multiple rows exceeding 100 m in length. In this work, we address a critical gap in the landscape of agricultural datasets by providing researchers with a comprehensive data collection and ground truth trajectories in different seasons, vegetation growth stages, terrain and weather conditions. The sequence paths with multiple runs and revisits will foster the development of sensor fusion, localization, mapping and place recognition solutions for agricultural fields. The dataset, the processing tools and the benchmarking results will be available at the dedicated webpage upon acceptance.

Paperid: 672, https://arxiv.org/pdf/2512.04332.pdf

Abstract:
Aligning generative diffusion models with human preferences via reinforcement learning (RL) is critical yet challenging. Most existing algorithms are often vulnerable to reward hacking, such as quality degradation, over-stylization, or reduced diversity. Our analysis demonstrates that this can be attributed to the inherent limitations of their regularization, which provides unreliable penalties. We introduce Data-regularized Diffusion Reinforcement Learning (DDRL), a novel framework that uses the forward KL divergence to anchor the policy to an off-policy data distribution. Theoretically, DDRL enables robust, unbiased integration of RL with standard diffusion training. Empirically, this translates into a simple yet effective algorithm that combines reward maximization with diffusion loss minimization. With over a million GPU hours of experiments and ten thousand double-blind human evaluations, we demonstrate on high-resolution video generation tasks that DDRL significantly improves rewards while alleviating the reward hacking seen in baselines, achieving the highest human preference and establishing a robust and scalable paradigm for diffusion post-training.

Paperid: 673, https://arxiv.org/pdf/2512.02293.pdf

Abstract:
We present VIGS-SLAM, a visual-inertial 3D Gaussian Splatting SLAM system that achieves robust real-time tracking and high-fidelity reconstruction. Although recent 3DGS-based SLAM methods achieve dense and photorealistic mapping, their purely visual design degrades under motion blur, low texture, and exposure variations. Our method tightly couples visual and inertial cues within a unified optimization framework, jointly refining camera poses, depths, and IMU states. It features robust IMU initialization, time-varying bias modeling, and loop closure with consistent Gaussian updates. Experiments on four challenging datasets demonstrate our superiority over state-of-the-art methods. Project page: https://vigs-slam.github.io

Paperid: 674, https://arxiv.org/pdf/2512.00947.pdf

Abstract:
To migrate the remarkable successes of Large Language Models (LLMs), the community has made numerous efforts to generalize them to the table reasoning tasks for the widely deployed tabular data. Despite that, in this work, by showing a probing experiment on our proposed StructQA benchmark, we postulate that even the most advanced LLMs (such as GPTs) may still fall short of coping with tabular data. More specifically, the current scheme often simply relies on serializing the tabular data, together with the meta information, then inputting them through the LLMs. We argue that the loss of structural information is the root of this shortcoming. In this work, we further propose TAMO, which bears an ideology to treat the tables as an independent modality integrated with the text tokens. The resulting model in TAMO is a multimodal framework consisting of a hypergraph neural network as the global table encoder seamlessly integrated with the mainstream LLM. Empirical results on various benchmarking datasets, including HiTab, WikiTQ, WikiSQL, FeTaQA, and StructQA, have demonstrated significant improvements on generalization with an average relative gain of 42.65%.

Paperid: 675, https://arxiv.org/pdf/2512.00878.pdf

Abstract:
Low-Rank Adaptation (LoRA) is a widely adopted parameter-efficient fine-tuning (PEFT) method for Large Language Models (LLMs), but it still incurs notable overhead and suffers from parameter interference in complex datasets. While re- cent works decouple LoRA update matrices to exploit matrix-wise asymmetry, training costs remain high. We revisit LoRA from the perspective of inter-matrix and intra-layer parameter redundancy and propose Resource-Efficient Low-Rank Adaptation, EffiLoRA, a lightweight and generalizable approach for language, multimodal, and diffusion models. EffiLoRA employs a unified A matrix across all transformer layers and introduces a runtime selective B matrices up- date to dynamically trade-off the system resource budget and model performance. EffiLoRA consistently outperforms LoRA across diverse modalities, including commonsense reasoning, visual instruction tuning, and image generation, demon- strating improved efficiency and robustness.

Paperid: 676, https://arxiv.org/pdf/2512.06746.pdf

Abstract:
Vision Language Models (VLMs) are increasingly adopted for AI-generated images (AIGI) detection, yet converting VLMs into detectors requires substantial resource, while the resulting models still exhibit severe hallucinations. To probe the core issue, we conduct an empirical analysis and observe two characteristic behaviors: (i) fine-tuning VLMs on high-level semantic supervision strengthens semantic discrimination and well generalize to unseen data; (ii) fine-tuning VLMs on low-level pixel-artifact supervision yields poor transfer. We attribute VLMs' underperformance to task-model misalignment: semantics-oriented VLMs inherently lack sensitivity to fine-grained pixel artifacts, and semantically non-discriminative pixel artifacts thus exceeds their inductive biases. In contrast, we observe that conventional pixel-artifact detectors capture low-level pixel artifacts yet exhibit limited semantic awareness relative to VLMs, highlighting that distinct models are better matched to distinct tasks. In this paper, we formalize AIGI detection as two complementary tasks--semantic consistency checking and pixel-artifact detection--and show that neglecting either induces systematic blind spots. Guided by this view, we introduce the Task-Model Alignment principle and instantiate it as a two-branch detector, AlignGemini, comprising a VLM fine-tuned exclusively with pure semantic supervision and a pixel-artifact expert trained exclusively with pure pixel-artifact supervision. By enforcing orthogonal supervision on two simplified datasets, each branch trains to its strengths, producing complementary discrimination over semantic and pixel cues. On five in-the-wild benchmarks, AlignGemini delivers a +9.5 gain in average accuracy, supporting task-model alignment as an effective path to generalizable AIGI detection.

Paperid: 677, https://arxiv.org/pdf/2512.05753.pdf

Abstract:
The fast deployment of cognitive radar to counter jamming remains a critical challenge in modern warfare, where more efficient deployment leads to quicker detection of targets. Existing methods are primarily based on evolutionary algorithms, which are time-consuming and prone to falling into local optima. We tackle these drawbacks via the efficient inference of neural networks and propose a brand new framework: Fast Anti-Jamming Radar Deployment Algorithm (FARDA). We first model the radar deployment problem as an end-to-end task and design deep reinforcement learning algorithms to solve it, where we develop integrated neural modules to perceive heatmap information and a brand new reward format. Empirical results demonstrate that our method achieves coverage comparable to evolutionary algorithms while deploying radars approximately 7,000 times faster. Further ablation experiments confirm the necessity of each component of FARDA.

Paperid: 678, https://arxiv.org/pdf/2512.05115.pdf

Abstract:
Recent advances in illumination control extend image-based methods to video, yet still facing a trade-off between lighting fidelity and temporal consistency. Moving beyond relighting, a key step toward generative modeling of real-world scenes is the joint control of camera trajectory and illumination, since visual dynamics are inherently shaped by both geometry and lighting. To this end, we present Light-X, a video generation framework that enables controllable rendering from monocular videos with both viewpoint and illumination control. 1) We propose a disentangled design that decouples geometry and lighting signals: geometry and motion are captured via dynamic point clouds projected along user-defined camera trajectories, while illumination cues are provided by a relit frame consistently projected into the same geometry. These explicit, fine-grained cues enable effective disentanglement and guide high-quality illumination. 2) To address the lack of paired multi-view and multi-illumination videos, we introduce Light-Syn, a degradation-based pipeline with inverse-mapping that synthesizes training pairs from in-the-wild monocular footage. This strategy yields a dataset covering static, dynamic, and AI-generated scenes, ensuring robust training. Extensive experiments show that Light-X outperforms baseline methods in joint camera-illumination control and surpasses prior video relighting methods under both text- and background-conditioned settings.

Paperid: 679, https://arxiv.org/pdf/2512.04733.pdf

Abstract:
End-to-end autonomous driving (AD) systems increasingly adopt vision-language-action (VLA) models, yet they typically ignore the passenger's emotional state, which is central to comfort and AD acceptance. We introduce Open-Domain End-to-End (OD-E2E) autonomous driving, where an autonomous vehicle (AV) must interpret free-form natural-language commands, infer the emotion, and plan a physically feasible trajectory. We propose E3AD, an emotion-aware VLA framework that augments semantic understanding with two cognitively inspired components: a continuous Valenc-Arousal-Dominance (VAD) emotion model that captures tone and urgency from language, and a dual-pathway spatial reasoning module that fuses egocentric and allocentric views for human-like spatial cognition. A consistency-oriented training scheme, combining modality pretraining with preference-based alignment, further enforces coherence between emotional intent and driving actions. Across real-world datasets, E3AD improves visual grounding and waypoint planning and achieves state-of-the-art (SOTA) VAD correlation for emotion estimation. These results show that injecting emotion into VLA-style driving yields more human-aligned grounding, planning, and human-centric feedback.

Paperid: 680, https://arxiv.org/pdf/2512.03570.pdf

Abstract:
Wireless sensor networks (WSNs) are employed across a wide range of industrial applications where ultra-low power consumption is a critical prerequisite. At the same time, these systems must maintain a certain level of determinism to ensure reliable and predictable operation. In this view, time slotted channel hopping (TSCH) is a communication technology that meets both conditions, making it an attractive option for its usage in industrial WSNs. This work proposes the use of machine learning to learn the traffic pattern generated in networks based on the TSCH protocol, in order to turn nodes into a deep sleep state when no transmission is planned and thus to improve the energy efficiency of the WSN. The ability of machine learning models to make good predictions at different network levels in a typical tree network topology was analyzed in depth, showing how their capabilities degrade while approaching the root of the tree. The application of these models on simulated data based on an accurate modeling of wireless sensor nodes indicates that the investigated algorithms can be suitably used to further and substantially reduce the power consumption of a TSCH network.

Paperid: 681, https://arxiv.org/pdf/2512.03569.pdf

Abstract:
Wireless communication is increasingly used in industrial environments, since it supports mobility of interconnected devices. Among the transmission technologies operating in unlicensed bands available to this purpose, Wi-Fi is certainly one of the most interesting, because of its high performance and the relatively low deployment costs. Unfortunately, its dependability is often deemed unsuitable for real-time control systems. In this paper, the use of parallel redundancy is evaluated from a quantitative viewpoint, by considering a number of performance indices that are relevant for soft real-time applications. Analysis is carried out on a large dataset acquired from a real setup, to provide realistic insights on the advantages this kind of approaches can provide. As will be seen, deferred parallel redundancy provides clear advantages in terms of the worst-case transmission latency, at limited costs concerning the amount of consumed spectrum. Hence, it can be practically exploited every time a wireless connection is included in a control loop.

Paperid: 682, https://arxiv.org/pdf/2512.03454.pdf

Abstract:
Interpreting natural-language commands to localize target objects is critical for autonomous driving (AD). Existing visual grounding (VG) methods for autonomous vehicles (AVs) typically struggle with ambiguous, context-dependent instructions, as they lack reasoning over 3D spatial relations and anticipated scene evolution. Grounded in the principles of world models, we propose ThinkDeeper, a framework that reasons about future spatial states before making grounding decisions. At its core is a Spatial-Aware World Model (SA-WM) that learns to reason ahead by distilling the current scene into a command-aware latent state and rolling out a sequence of future latent states, providing forward-looking cues for disambiguation. Complementing this, a hypergraph-guided decoder then hierarchically fuses these states with the multimodal input, capturing higher-order spatial dependencies for robust localization. In addition, we present DrivePilot, a multi-source VG dataset in AD, featuring semantic annotations generated by a Retrieval-Augmented Generation (RAG) and Chain-of-Thought (CoT)-prompted LLM pipeline. Extensive evaluations on six benchmarks, ThinkDeeper ranks #1 on the Talk2Car leaderboard and surpasses state-of-the-art baselines on DrivePilot, MoCAD, and RefCOCO/+/g benchmarks. Notably, it shows strong robustness and efficiency in challenging scenes (long-text, multi-agent, ambiguity) and retains superior performance even when trained on 50% of the data.

Paperid: 683, https://arxiv.org/pdf/2512.02843.pdf

Abstract:
Scalability is a major challenge in non-geostationary orbit (NGSO) satellite networks due to the massive number of ground users sharing the limited sub-6 GHz spectrum. Using K- and higher bands is a promising alternative to increase the accessible bandwidth, but these bands are subject to significant atmospheric attenuation, notably rainfall, which can lead to degraded performance and link outages. We present an integrated sensing and communications (ISAC)-powered framework for resilient and efficient operation of multi-band satellite networks. It is based on distributed mechanisms for atmospheric sensing, cell-to-satellite matching, and resource allocation (RA) in a 5G Non-Terrestrial Network (NTN) wide-area scenario with quasi-Earth fixed cells and a beam hopping mechanism. Results with a multi-layer multi-band constellation with satellites operating in the S- and K-bands demonstrate the benefits of our framework for ISAC-powered multi-band systems, which achieves 73% higher throughput per user when compared to single S- and K-band systems.

Paperid: 684, https://arxiv.org/pdf/2512.02455.pdf

Abstract:
Wi-Fi is currently considered one of the most promising solutions for interconnecting mobile equipment (e.g., autonomous mobile robots and active exoskeletons) in industrial environments. However, relability requirements imposed by the industrial context, such as ensuring bounded transmission latency, are a major challenge for over-the-air communication. One of the aspects of Wi-Fi technology that greatly affects the probability of a packet reaching its destination is the selection of the appropriate transmission rate. Rate adaptation algorithms are in charge of this operation, but their design and implementation are not regulated by the IEEE 802.11 standard. One of the most popular solutions, available as open source, is Minstrel, which is the default choice for the Linux Kernel. In this paper, Minstrel performance is evaluated for both static and mobility scenarios. Our analysis focuses on metrics of interest for industrial contexts, i.e., latency and packet loss ratio, and serves as a preliminary evaluation for the future development of enhanced rate adaptation algorithms based on centralized digital twins.

Paperid: 685, https://arxiv.org/pdf/2512.02454.pdf

Abstract:
Precise clock synchronization protocols are increasingly used to ensure that all the nodes in a network share the very same time base. They enable several mechanisms aimed at improving determinism at both the application and communication levels, which makes them highly relevant to industrial environments. Reference Broadcast Infrastructure Synchronization (RBIS) is a solution specifically conceived for Wi-Fi that exploits existing beacons and can run on commercial devices. In this paper, an evolution of RBIS is presented, we call DOMINO, whose coverage area is much larger than the single Wi-Fi infrastructure network, potentially including the whole plant. In particular, wireless stations that can see more than one access point at the same time behave as boundary clocks and propagate the reference time across overlapping networks.

Paperid: 686, https://arxiv.org/pdf/2512.01818.pdf

Abstract:
Deep neural networks suffer from catastrophic forgetting, where performance on previous tasks degrades after training on a new task. This issue arises due to the model's tendency to overwrite previously acquired knowledge with new information. We present a novel approach to address this challenge, focusing on the intersection of memory-based methods and regularization approaches. We formulate a regularization strategy, termed Information Maximization (IM) regularizer, for memory-based continual learning methods, which is based exclusively on the expected label distribution, thus making it class-agnostic. As a consequence, IM regularizer can be directly integrated into various rehearsal-based continual learning methods, reducing forgetting and favoring faster convergence. Our empirical validation shows that, across datasets and regardless of the number of tasks, our proposed regularization strategy consistently improves baseline performance at the expense of a minimal computational overhead. The lightweight nature of IM ensures that it remains a practical and scalable solution, making it applicable to real-world continual learning scenarios where efficiency is paramount. Finally, we demonstrate the data-agnostic nature of our regularizer by applying it to video data, which presents additional challenges due to its temporal structure and higher memory requirements. Despite the significant domain gap, our experiments show that IM regularizer also improves the performance of video continual learning methods.

Paperid: 687, https://arxiv.org/pdf/2512.01228.pdf

Abstract:
Achieving optimality and adversarial robustness in deep reinforcement learning has long been regarded as conflicting goals. Nonetheless, recent theoretical insights presented in CAR suggest a potential alignment, raising the important question of how to realize this in practice. This paper first identifies a key gap between theory and practice by comparing standard policy optimization (SPO) and adversarially robust policy optimization (ARPO). Although they share theoretical consistency, a fundamental tension between robustness and optimality arises in practical policy gradient methods. SPO tends toward convergence to vulnerable first-order stationary policies (FOSPs) with strong natural performance, whereas ARPO typically favors more robust FOSPs at the expense of reduced returns. Furthermore, we attribute this tradeoff to the reshaping effect of the strongest adversary in ARPO, which significantly complicates the global landscape by inducing deceptive sticky FOSPs. This improves robustness but makes navigation more challenging. To alleviate this, we develop the BARPO, a bilevel framework unifying SPO and ARPO by modulating adversary strength, thereby facilitating navigability while preserving global optima. Extensive empirical results demonstrate that BARPO consistently outperforms vanilla ARPO, providing a practical approach to reconcile theoretical and empirical performance.

Paperid: 688, https://arxiv.org/pdf/2512.00877.pdf

Abstract:
3D Gaussian Splatting (3DGS) has emerged as a revolutionary 3D representation. However, its substantial data size poses a major barrier to widespread adoption. While feed-forward 3DGS compression offers a practical alternative to costly per-scene per-train compressors, existing methods struggle to model long-range spatial dependencies, due to the limited receptive field of transform coding networks and the inadequate context capacity in entropy models. In this work, we propose a novel feed-forward 3DGS compression framework that effectively models long-range correlations to enable highly compact and generalizable 3D representations. Central to our approach is a large-scale context structure that comprises thousands of Gaussians based on Morton serialization. We then design a fine-grained space-channel auto-regressive entropy model to fully leverage this expansive context. Furthermore, we develop an attention-based transform coding model to extract informative latent priors by aggregating features from a wide range of neighboring Gaussians. Our method yields a $20\times$ compression ratio for 3DGS in a feed-forward inference and achieves state-of-the-art performance among generalizable codecs.

Paperid: 689, https://arxiv.org/pdf/2512.00475.pdf

Abstract:
Generic Event Boundary Detection (GEBD) aims to identify moments in videos that humans perceive as event boundaries. This paper proposes a novel method for addressing this task, called Structured Context Learning, which introduces the Structured Partition of Sequence (SPoS) to provide a structured context for learning temporal information. Our approach is end-to-end trainable and flexible, not restricted to specific temporal models like GRU, LSTM, and Transformers. This flexibility enables our method to achieve a better speed-accuracy trade-off. Specifically, we apply SPoS to partition the input frame sequence and provide a structured context for the subsequent temporal model. Notably, SPoS's overall computational complexity is linear with respect to the video length. We next calculate group similarities to capture differences between frames, and a lightweight fully convolutional network is utilized to determine the event boundaries based on the grouped similarity maps. To remedy the ambiguities of boundary annotations, we adapt the Gaussian kernel to preprocess the ground-truth event boundaries. Our proposed method has been extensively evaluated on the challenging Kinetics-GEBD, TAPOS, and shot transition detection datasets, demonstrating its superiority over existing state-of-the-art methods.

Paperid: 690, https://arxiv.org/pdf/2512.00287.pdf

Abstract:
Existing appliance assets suffer from poor rendering, incomplete mechanisms, and misalignment with manuals, leading to simulation-reality gaps that hinder appliance manipulation development. In this work, we introduce the RealAppliance dataset, comprising 100 high-fidelity appliances with complete physical, electronic mechanisms, and program logic aligned with their manuals. Based on these assets, we propose the RealAppliance-Bench benchmark, which evaluates multimodal large language models and embodied manipulation planning models across key tasks in appliance manipulation planning: manual page retrieval, appliance part grounding, open-loop manipulation planning, and closed-loop planning adjustment. Our analysis of model performances on RealAppliance-Bench provides insights for advancing appliance manipulation research

Paperid: 691, https://arxiv.org/pdf/2512.00221.pdf

Abstract:
Ensuring reliable and predictable communications is one of the main goals in modern industrial systems that rely on Wi-Fi networks, especially in scenarios where continuity of operation and low latency are required. In these contexts, the ability to predict changes in wireless channel quality can enable adaptive strategies and significantly improve system robustness. This contribution focuses on the prediction of the Frame Delivery Ratio (FDR), a key metric that represents the percentage of successful transmissions, starting from time sequences of binary outcomes (success/failure) collected in a real scenario. The analysis focuses on two models of deep learning: a Convolutional Neural Network (CNN) and a Long Short-Term Memory network (LSTM), both selected for their ability to predict the outcome of time sequences. Models are compared in terms of prediction accuracy and computational complexity, with the aim of evaluating their applicability to systems with limited resources. Preliminary results show that both models are able to predict the evolution of the FDR with good accuracy, even from minimal information (a single binary sequence). In particular, CNN shows a significantly lower inference latency, with a marginal loss in accuracy compared to LSTM.

Paperid: 693, https://arxiv.org/pdf/2512.07363.pdf

Abstract:
Precise temporal and spatial alignment is critical in collaborative Augmented Reality (AR) where users rely on shared visual information to coordinate actions. System latency and object misalignment can disrupt communication, reduce task efficiency, and negatively impact the overall user experience. While previous research has primarily focused on individual AR interactions, the impact of these inconsistencies on collaboration remains underexplored. This article investigates how user experience and task load are affected by object misalignment and time delay in a shared AR space. To examine these factors, we conducted an experiment with 32 participants, organized into 16 pairs, who collaboratively completed a spatial placement task. Within each condition, both participants alternated roles, taking turns as the leader-providing verbal placement instructions-and the builder-executing the placement. Six conditions were tested, manipulating object alignment (perfectly aligned vs. randomly misaligned) and time delay (0s, 0.1s, 0.4s). The misalignment was applied randomly to each virtual object with a shift of +-20 cm on every axis to create a clear distinction in spatial perception. User experience and task load were assessed to evaluate how these factors influence collaboration and interaction in AR environments. Results showed that spatial misalignment significantly increased perceived workload (NASA-TLX) and lowered user ratings in Pragmatic quality and Attractiveness (UEQ), while time delay had a more limited effect. These findings highlight the critical role of spatial accuracy in maintaining collaboration quality in AR.

Paperid: 694, https://arxiv.org/pdf/2512.07357.pdf

Abstract:
The usage of virtual avatars in healthcare applications has become widely popular; however, certain critical aspects, such as social distancing and avatar size, remain insufficiently explored. This research investigates user experience and preferences when interacting with a healthcare application utilizing virtual avatars displayed in different sizes. For our study, we had 23 participants interacting with five different avatars (a human-size avatar followed by four smaller avatars in a randomized order) varying in size, projected on a wall in front of them. The avatars were fully integrated with an artificial intelligence chatbot to make them conversational. Users were asked to rate the usability of the system after interacting with each avatar and complete a survey regarding trust and an additional questionnaire on social presence. The results of this study show that avatar size significantly influences the perceived attractiveness and perspicuity, with the medium-sized avatars receiving the highest ratings. Social presence correlated strongly with stimulation and attractiveness, suggesting that an avatar's visual appeal and interactivity influenced user engagement more than its physical size. Additionally, we observed a tendency for gender-specific differences on some of the UEQ+ scales, with male participants tending to prefer human-sized representations, while female participants slightly favored smaller avatars. These findings highlight the importance of avatar design and representation in optimizing user experience and trust in virtual healthcare environments.

Paperid: 695, https://arxiv.org/pdf/2512.06802.pdf

Abstract:
The rapid development of generative models has significantly advanced image and video applications. Among these, video creation, aimed at generating videos under various conditions, has gained substantial attention. However, existing video creation models either focus solely on a few specific conditions or suffer from excessively long generation times due to complex model inference, making them impractical for real-world applications. To mitigate these issues, we propose an efficient unified video creation model, named VDOT. Concretely, we model the training process with the distribution matching distillation (DMD) paradigm. Instead of using the Kullback-Leibler (KL) minimization, we additionally employ a novel computational optimal transport (OT) technique to optimize the discrepancy between the real and fake score distributions. The OT distance inherently imposes geometric constraints, mitigating potential zero-forcing or gradient collapse issues that may arise during KL-based distillation within the few-step generation scenario, and thus, enhances the efficiency and stability of the distillation process. Further, we integrate a discriminator to enable the model to perceive real video data, thereby enhancing the quality of generated videos. To support training unified video creation models, we propose a fully automated pipeline for video data annotation and filtering that accommodates multiple video creation tasks. Meanwhile, we curate a unified testing benchmark, UVCBench, to standardize evaluation. Experiments demonstrate that our 4-step VDOT outperforms or matches other baselines with 100 denoising steps.

Paperid: 696, https://arxiv.org/pdf/2512.06774.pdf

Abstract:
3D Gaussian Splatting (3DGS) has enabled the creation of digital assets and downstream applications, underscoring the need for robust copyright protection via digital watermarking. However, existing 3DGS watermarking methods remain highly vulnerable to diffusion-based editing, which can easily erase embedded provenance. This challenge highlights the urgent need for 3DGS watermarking techniques that are intrinsically resilient to diffusion-based editing. In this paper, we introduce RDSplat, a Robust watermarking paradigm against Diffusion editing for 3D Gaussian Splatting. RDSplat embeds watermarks into 3DGS components that diffusion-based editing inherently preserve, achieved through (i) proactively targeting low-frequency Gaussians and (ii) adversarial training with a diffusion proxy. Specifically, we introduce a multi-domain framework that operates natively in 3DGS space and embeds watermarks into diffusion-editing-preserved low-frequency Gaussians via coordinated covariance regularization and 2D filtering. In addition, we exploit the low-pass filtering behavior of diffusion-based editing by using Gaussian blur as an efficient training surrogate, enabling adversarial fine-tuning that further enhances watermark robustness against diffusion-based editing. Empirically, comprehensive quantitative and qualitative evaluations on three benchmark datasets demonstrate that RDSplat not only maintains superior robustness under diffusion-based editing, but also preserves watermark invisibility, achieving state-of-the-art performance.

Paperid: 697, https://arxiv.org/pdf/2512.06261.pdf

Abstract:
Generating safe, kinodynamically feasible, and optimal trajectories for complex robotic systems is a central challenge in robotics. This paper presents Safe Model Predictive Diffusion (Safe MPD), a training-free diffusion planner that unifies a model-based diffusion framework with a safety shield to generate trajectories that are both kinodynamically feasible and safe by construction. By enforcing feasibility and safety on all samples during the denoising process, our method avoids the common pitfalls of post-processing corrections, such as computational intractability and loss of feasibility. We validate our approach on challenging non-convex planning problems, including kinematic and acceleration-controlled tractor-trailer systems. The results show that it substantially outperforms existing safety strategies in success rate and safety, while achieving sub-second computation times.

Paperid: 698, https://arxiv.org/pdf/2512.04568.pdf

Abstract:
Inspired by traditional handmade crafts, where a person improvises assemblies based on the available objects, we formally introduce the Craft Assembly Task. It is a robotic assembly task that involves building an accurate representation of a given target object using the available objects, which do not directly correspond to its parts. In this work, we focus on selecting the subset of available objects for the final craft, when the given input is an RGB image of the target in the wild. We use a mask segmentation neural network to identify visible parts, followed by retrieving labeled template meshes. These meshes undergo pose optimization to determine the most suitable template. Then, we propose to simplify the parts of the transformed template mesh to primitive shapes like cuboids or cylinders. Finally, we design a search algorithm to find correspondences in the scene based on local and global proportions. We develop baselines for comparison that consider all possible combinations, and choose the highest scoring combination for common metrics used in foreground maps and mask accuracy. Our approach achieves comparable results to the baselines for two different scenes, and we show qualitative results for an implementation in a real-world scenario.

Paperid: 699, https://arxiv.org/pdf/2512.04421.pdf

Abstract:
Ray tracing 3D Gaussian particles enables realistic effects such as depth of field, refractions, and flexible camera modeling for novel-view synthesis. However, existing methods trace Gaussians through proxy geometry, which requires constructing complex intermediate meshes and performing costly intersection tests. This limitation arises because Gaussian-based particles are not well suited as unified primitives for both ray tracing and rasterization. In this work, we propose a differentiable triangle-based ray tracing pipeline that directly treats triangles as rendering primitives without relying on any proxy geometry. Our results show that the proposed method achieves significantly higher rendering quality than existing ray tracing approaches while maintaining real-time rendering performance. Moreover, our pipeline can directly render triangles optimized by the rasterization-based method Triangle Splatting, thus unifying the primitives used in novel-view synthesis.

Paperid: 700, https://arxiv.org/pdf/2512.03724.pdf

Abstract:
The Vision-Language-Action (VLA) models have demonstrated remarkable performance on embodied tasks and shown promising potential for real-world applications. However, current VLAs still struggle to produce consistent and precise target-oriented actions, as they often generate redundant or unstable motions along trajectories, limiting their applicability in time-sensitive scenarios.In this work, we attribute these redundant actions to the spatially uniform perception field of existing VLAs, which causes them to be distracted by target-irrelevant objects, especially in complex environments.To address this issue, we propose an efficient PosA-VLA framework that anchors visual attention via pose-conditioned supervision, consistently guiding the model's perception toward task-relevant regions. The pose-conditioned anchor attention mechanism enables the model to better align instruction semantics with actionable visual cues, thereby improving action generation precision and efficiency. Moreover, our framework adopts a lightweight architecture and requires no auxiliary perception modules (e.g., segmentation or grounding networks), ensuring efficient inference. Extensive experiments verify that our method executes embodied tasks with precise and time-efficient behavior across diverse robotic manipulation benchmarks and shows robust generalization in a variety of challenging environments.

Paperid: 701, https://arxiv.org/pdf/2512.03336.pdf

Abstract:
Federated Learning (FL) is plagued by two key challenges: high communication overhead and performance collapse on heterogeneous (non-IID) data. Analytic FL (AFL) provides a single-round, data distribution invariant solution, but is limited to linear models. Subsequent non-linear approaches, like DeepAFL, regain accuracy but sacrifice the single-round benefit. In this work, we break this trade-off. We propose SAFLe, a framework that achieves scalable non-linear expressivity by introducing a structured head of bucketed features and sparse, grouped embeddings. We prove this non-linear architecture is mathematically equivalent to a high-dimensional linear regression. This key equivalence allows SAFLe to be solved with AFL's single-shot, invariant aggregation law. Empirically, SAFLe establishes a new state-of-the-art for analytic FL, significantly outperforming both linear AFL and multi-round DeepAFL in accuracy across all benchmarks, demonstrating a highly efficient and scalable solution for federated vision.

Paperid: 702, https://arxiv.org/pdf/2512.02902.pdf

Abstract:
Vision-language-action (VLA) models achieve strong in-distribution performance but degrade sharply under novel camera viewpoints and visual perturbations. We show that this brittleness primarily arises from misalignment in Spatial Modeling, rather than Physical Modeling. To address this, we propose a one-shot adaptation framework that recalibrates visual representations through lightweight, learnable updates. Our first method, Feature Token Modulation (FTM), applies a global affine transformation to visual tokens and improves Libero viewpoint accuracy from 48.5% to 87.1% with only 4K parameters. Building on this, Feature Linear Adaptation (FLA) introduces low-rank updates to the ViT encoder, achieving 90.8% success with 4.7M parameters -- matching LoRA-scale finetuning at far lower cost. Together, these results reveal substantial untapped robustness in pretrained VLA models and demonstrate that targeted, minimal visual adaptation is sufficient to restore viewpoint generalization.

Paperid: 703, https://arxiv.org/pdf/2512.02069.pdf

Abstract:
This paper presents LLMBugScanner, a large language model (LLM) based framework for smart contract vulnerability detection using fine-tuning and ensemble learning. Smart contract auditing presents several challenges for LLMs: different pretrained models exhibit varying reasoning abilities, and no single model performs consistently well across all vulnerability types or contract structures. These limitations persist even after fine-tuning individual LLMs. To address these challenges, LLMBugScanner combines domain knowledge adaptation with ensemble reasoning to improve robustness and generalization. Through domain knowledge adaptation, we fine-tune LLMs on complementary datasets to capture both general code semantics and instruction-guided vulnerability reasoning, using parameter-efficient tuning to reduce computational cost. Through ensemble reasoning, we leverage the complementary strengths of multiple LLMs and apply a consensus-based conflict resolution strategy to produce more reliable vulnerability assessments. We conduct extensive experiments across multiple popular LLMs and compare LLMBugScanner with both pretrained and fine-tuned individual models. Results show that LLMBugScanner achieves consistent accuracy improvements and stronger generalization, demonstrating that it provides a principled, cost-effective, and extensible framework for smart contract auditing.

Paperid: 704, https://arxiv.org/pdf/2512.02058.pdf

Abstract:
Recent advances enable Large Language Models (LLMs) to generate AI personas, yet their lack of deep contextual, cultural, and emotional understanding poses a significant limitation. This study quantitatively compared human responses with those of eight LLM-generated social personas (e.g., Male, Female, Muslim, Political Supporter) within a low-resource environment like Bangladesh, using culturally specific questions. Results show human responses significantly outperform all LLMs in answering questions, and across all matrices of persona perception, with particularly large gaps in empathy and credibility. Furthermore, LLM-generated content exhibited a systematic bias along the lines of the ``Pollyanna Principle'', scoring measurably higher in positive sentiment ($Φ_{avg} = 5.99$ for LLMs vs. $5.60$ for Humans). These findings suggest that LLM personas do not accurately reflect the authentic experience of real people in resource-scarce environments. It is essential to validate LLM personas against real-world human data to ensure their alignment and reliability before deploying them in social science research.

Paperid: 705, https://arxiv.org/pdf/2512.07498.pdf

Abstract:
Ensuring the authenticity of video content remains challenging as DeepFake generation becomes increasingly realistic and robust against detection. Most existing detectors implicitly assume temporally consistent and clean facial sequences, an assumption that rarely holds in real-world scenarios where compression artifacts, occlusions, and adversarial attacks destabilize face detection and often lead to invalid or misdetected faces. To address these challenges, we propose a Laplacian-Regularized Graph Convolutional Network (LR-GCN) that robustly detects DeepFakes from noisy or unordered face sequences, while being trained only on clean facial data. Our method constructs an Order-Free Temporal Graph Embedding (OF-TGE) that organizes frame-wise CNN features into an adaptive sparse graph based on semantic affinities. Unlike traditional methods constrained by strict temporal continuity, OF-TGE captures intrinsic feature consistency across frames, making it resilient to shuffled, missing, or heavily corrupted inputs. We further impose a dual-level sparsity mechanism on both graph structure and node features to suppress the influence of invalid faces. Crucially, we introduce an explicit Graph Laplacian Spectral Prior that acts as a high-pass operator in the graph spectral domain, highlighting structural anomalies and forgery artifacts, which are then consolidated by a low-pass GCN aggregation. This sequential design effectively realizes a task-driven spectral band-pass mechanism that suppresses background information and random noise while preserving manipulation cues. Extensive experiments on FF++, Celeb-DFv2, and DFDC demonstrate that LR-GCN achieves state-of-the-art performance and significantly improved robustness under severe global and local disruptions, including missing faces, occlusions, and adversarially perturbed face detections.

Paperid: 706, https://arxiv.org/pdf/2512.07436.pdf

Abstract:
Recent advances in large reasoning models (LRMs) have enabled agentic search systems to perform complex multi-step reasoning across multiple sources. However, most studies focus on general information retrieval and rarely explores vertical domains with unique challenges. In this work, we focus on local life services and introduce LocalSearchBench, which encompass diverse and complex business scenarios. Real-world queries in this domain are often ambiguous and require multi-hop reasoning across merchants and products, remaining challenging and not fully addressed. As the first comprehensive benchmark for agentic search in local life services, LocalSearchBench includes over 150,000 high-quality entries from various cities and business types. We construct 300 multi-hop QA tasks based on real user queries, challenging agents to understand questions and retrieve information in multiple steps. We also developed LocalPlayground, a unified environment integrating multiple tools for agent interaction. Experiments show that even state-of-the-art LRMs struggle on LocalSearchBench: the best model (DeepSeek-V3.1) achieves only 34.34% correctness, and most models have issues with completeness (average 77.33%) and faithfulness (average 61.99%). This highlights the need for specialized benchmarks and domain-specific agent training in local life services. Code, Benchmark, and Leaderboard are available at localsearchbench.github.io.

Paperid: 707, https://arxiv.org/pdf/2512.04446.pdf

Abstract:
Automating disassembly of critical components from end-of-life (EoL) desktops, such as high-value items like RAM modules and CPUs, as well as sensitive parts like hard disk drives, remains challenging due to the inherent variability and uncertainty of these products. Moreover, their disassembly requires sequential, precise, and dexterous operations, further increasing the complexity of automation. Current robotic disassembly processes are typically divided into several stages: perception, sequence planning, task planning, motion planning, and manipulation. Each stage requires explicit modeling, which limits generalization to unfamiliar scenarios. Recent development of vision-language-action (VLA) models has presented an end-to-end approach for general robotic manipulation tasks. Although VLAs have demonstrated promising performance on simple tasks, the feasibility of applying such models to complex disassembly remains largely unexplored. In this paper, we collected a customized dataset for robotic RAM and CPU disassembly and used it to fine-tune two well-established VLA approaches, OpenVLA and OpenVLA-OFT, as a case study. We divided the whole disassembly task into several small steps, and our preliminary experimental results indicate that the fine-tuned VLA models can faithfully complete multiple early steps but struggle with certain critical subtasks, leading to task failure. However, we observed that a simple hybrid strategy that combines VLA with a rule-based controller can successfully perform the entire disassembly operation. These findings highlight the current limitations of VLA models in handling the dexterity and precision required for robotic EoL product disassembly. By offering a detailed analysis of the observed results, this study provides insights that may inform future research to address current challenges and advance end-to-end robotic automated disassembly.

Paperid: 708, https://arxiv.org/pdf/2512.04415.pdf

Abstract:
Physical feasibility in 3D bin packing is a key requirement in modern industrial logistics and robotic automation. With the growing adoption of industrial automation, online bin packing has gained increasing attention. However, inconsistencies in problem settings, test datasets, and evaluation metrics have hindered progress in the field, and there is a lack of a comprehensive benchmarking system. Direct testing on real hardware is costly, and building a realistic simulation environment is also challenging. To address these limitations, we introduce RoboBPP, a benchmarking system designed for robotic online bin packing. RoboBPP integrates a physics-based simulator to assess physical feasibility. In our simulation environment, we introduce a robotic arm and boxes at real-world scales to replicate real industrial packing workflows. By simulating conditions that arise in real industrial applications, we ensure that evaluated algorithms are practically deployable. In addition, prior studies often rely on synthetic datasets whose distributions differ from real-world industrial data. To address this issue, we collect three datasets from real industrial workflows, including assembly-line production, logistics packing, and furniture manufacturing. The benchmark comprises three carefully designed test settings and extends existing evaluation metrics with new metrics for structural stability and operational safety. We design a scoring system and derive a range of insights from the evaluation results. RoboBPP is fully open-source and is equipped with visualization tools and an online leaderboard, providing a reproducible and extensible foundation for future research and industrial applications (https://robot-bin-packing-benchmark.github.io).

Paperid: 709, https://arxiv.org/pdf/2512.03444.pdf

Abstract:
Deep learning methods have significantly enhanced motion planning for robotic manipulators by leveraging prior experiences within planning datasets. However, state-of-the-art neural motion planners are primarily trained on small datasets collected in manually generated workspaces, limiting their generalizability to out-of-distribution scenarios. Additionally, these planners often rely on monolithic network architectures that struggle to encode critical planning information. To address these challenges, we introduce Motion Policy with Dataset Synthesis powered by large language models (LLMs) and Fusion Action-Chunking Transformers (PerFACT), which incorporates two key components. Firstly, a novel LLM-powered workspace generation method, MotionGeneralizer, enables large-scale planning data collection by producing a diverse set of semantically feasible workspaces. Secondly, we introduce Fusion Motion Policy Networks (MpiNetsFusion), a generalist neural motion planner that uses a fusion action-chunking transformer to better encode planning signals and attend to multiple feature modalities. Leveraging MotionGeneralizer, we collect 3.5M trajectories to train and evaluate MpiNetsFusion against state-of-the-art planners, which shows that the proposed MpiNetsFusion can plan several times faster on the evaluated tasks.

Paperid: 710, https://arxiv.org/pdf/2512.02811.pdf

Abstract:
The notion of synthetic molecular communication (MC) refers to the transmission of information via signaling molecules and is foreseen to enable innovative medical applications in the human cardiovascular system (CVS). Crucially, the design of such applications requires accurate and experimentally validated channel models that characterize the propagation of signaling molecules, not just in individual blood vessels, but in complex vessel networks (VNs), as prevalent in the CVS. However, experimentally validated models for MC in VNs remain scarce. To address this gap, we propose a novel channel model for MC in complex VN topologies, which captures molecular transport via advection, molecular and turbulent diffusion, as well as adsorption and desorption at the vessel walls. We specialize this model for superparamagnetic iron-oxide nanoparticles (SPIONs) as signaling molecules by introducing a new receiver (RX) model for planar coil inductive sensors, enabling end-to-end experimental validation with a dedicated SPION testbed. Validation covers a range of channel topologies, from single-vessel topologies to branched VNs with multiple paths between transmitter (TX) and RX. Additionally, to quantify how the VN topology impacts signal quality, and inspired by multi-path propagation models in conventional wireless communications, we introduce two metrics, namely molecule delay and multi-path spread. We show that these metrics link the VN structure to molecule dispersion induced by the VN and mediately to the resulting signal-to-noise ratio (SNR) at the RX. The proposed VN structure-SNR link is validated experimentally, demonstrating that the proposed framework can support tasks such as optimal sensor placement in the CVS or the identification of suitable testbed topologies for specific SNR requirements. All experimental data are openly available on Zenodo.

Paperid: 711, https://arxiv.org/pdf/2512.01582.pdf

Abstract:
In this paper, we introduce RoleMotion, a large-scale human motion dataset that encompasses a wealth of role-playing and functional motion data tailored to fit various specific scenes. Existing text datasets are mainly constructed decentrally as amalgamation of assorted subsets that their data are nonfunctional and isolated to work together to cover social activities in various scenes. Also, the quality of motion data is inconsistent, and textual annotation lacks fine-grained details in these datasets. In contrast, RoleMotion is meticulously designed and collected with a particular focus on scenes and roles. The dataset features 25 classic scenes, 110 functional roles, over 500 behaviors, and 10296 high-quality human motion sequences of body and hands, annotated with 27831 fine-grained text descriptions. We build an evaluator stronger than existing counterparts, prove its reliability, and evaluate various text-to-motion methods on our dataset. Finally, we explore the interplay of motion generation of body and hands. Experimental results demonstrate the high-quality and functionality of our dataset on text-driven whole-body generation.

Paperid: 712, https://arxiv.org/pdf/2512.00904.pdf

Abstract:
Image clustering is a classic problem in computer vision, which categorizes images into different groups. Recent studies utilize nouns as external semantic knowledge to improve clus- tering performance. However, these methods often overlook the inherent ambiguity of nouns, which can distort semantic representations and degrade clustering quality. To address this issue, we propose a hierarChical semAntic alignmEnt method for image clustering, dubbed CAE, which improves cluster- ing performance in a training-free manner. In our approach, we incorporate two complementary types of textual seman- tics: caption-level descriptions, which convey fine-grained attributes of image content, and noun-level concepts, which represent high-level object categories. We first select relevant nouns from WordNet and descriptions from caption datasets to construct a semantic space aligned with image features. Then, we align image features with selected nouns and captions via optimal transport to obtain a more discriminative semantic space. Finally, we combine the enhanced semantic and image features to perform clustering. Extensive experiments across 8 datasets demonstrate the effectiveness of our method, notably surpassing the state-of-the-art training-free approach with a 4.2% improvement in accuracy and a 2.9% improvement in adjusted rand index (ARI) on the ImageNet-1K dataset.

Paperid: 713, https://arxiv.org/pdf/2512.00229.pdf

Abstract:
Deep neural networks often struggle to recognize when an input lies outside their training experience, leading to unreliable and overconfident predictions. Building dependable machine learning systems therefore requires methods that can both estimate predictive \textit{uncertainty} and detect \textit{out-of-distribution (OOD)} samples in a unified manner. In this paper, we propose \textbf{TIE: a Training--Inversion--Exclusion} framework for visually interpretable and uncertainty-guided anomaly detection that jointly addresses these challenges through iterative refinement. TIE extends a standard $n$-class classifier to an $(n+1)$-class model by introducing a garbage class initialized with Gaussian noise to represent outlier inputs. Within each epoch, TIE performs a closed-loop process of \textit{training, inversion, and exclusion}, where highly uncertain inverted samples reconstructed from the just-trained classifier are excluded into the garbage class. Over successive iterations, the inverted samples transition from noisy artifacts into visually coherent class prototypes, providing transparent insight into how the model organizes its learned manifolds. During inference, TIE rejects OOD inputs by either directly mapping them to the garbage class or producing low-confidence, uncertain misclassifications within the in-distribution classes that are easily separable, all without relying on external OOD datasets. A comprehensive threshold-based evaluation using multiple OOD metrics and performance measures such as \textit{AUROC}, \textit{AUPR}, and \textit{FPR@95\%TPR} demonstrates that TIE offers a unified and interpretable framework for robust anomaly detection and calibrated uncertainty estimation (UE) achieving near-perfect OOD detection with \textbf{$\!\approx\!$ 0 FPR@95\%TPR} when trained on MNIST or FashionMNIST and tested against diverse unseen datasets.

Paperid: 714, https://arxiv.org/pdf/2512.00107.pdf

Abstract:
Identifying central entities and interactions is a fundamental problem in network science. While well-studied for graphs (pairwise relations), many biological and social systems exhibit higher-order interactions best modeled by hypergraphs. This has led to a proliferation of specialized hypergraph centrality measures, but the field remains fragmented and lacks a unifying framework. This paper addresses this gap by providing the first systematic survey of 39 distinct measures. We introduce a novel taxonomy classifying them as: (1) structural (topology-based), (2) functional (impact on system dynamics), or (3) contextual (incorporating external features). We also present an experimental assessment comparing their empirical similarity and computation time. Finally, we discuss applications, establishing a coherent roadmap for future research in this area.

Paperid: 715, https://arxiv.org/pdf/2512.06743.pdf

Abstract:
Road network data can provide rich information about cities and thus become the base for various urban research. However, processing large volume world-wide road network data requires intensive computing resources and the processed results might be different to be unified for testing downstream tasks. Therefore, in this paper, we process the OpenStreetMap data via a distributed computing of 5,000 cores on cloud services and release a structured world-wide 1-billion-vertex road network graph dataset with high accessibility (opensource and downloadable to the whole world) and usability (open-box graph structure and easy spatial query interface). To demonstrate how this dataset can be utilized easily, we present three illustrative use cases, including traffic prediction, city boundary detection and traffic policy control, and conduct extensive experiments for these three tasks. (1) For the well-investigated traffic prediction tasks, we release a new benchmark with 31 cities (traffic data processed and combined with our released OSM+ road network dataset), to provide much larger spatial coverage and more comprehensive evaluation of compared algorithms than the previously frequently-used datasets. This new benchmark will push the algorithms on their scalability from hundreds of road network intersections to thousands of intersections. (2) While for the more advanced traffic policy control task which requires interaction with the road network, we release a new 6 city datasets with much larger scale than the previous datasets. This brings new challenge for thousand-scale multi-agent coordination. (3) Along with the OSM+ dataset, the release of data converters facilitates the integration of multimodal spatial-temporal data for geospatial foundation model training, thereby expediting the process of uncovering compelling scientific insights. PVLDB Reference Forma

Paperid: 716, https://arxiv.org/pdf/2512.05094.pdf

Abstract:
Video generation models are rapidly improving in their ability to synthesize human actions in novel contexts, holding the potential to serve as high-level planners for contextual robot control. To realize this potential, a key research question remains open: how can a humanoid execute the human actions from generated videos in a zero-shot manner? This challenge arises because generated videos are often noisy and exhibit morphological distortions that make direct imitation difficult compared to real video. To address this, we introduce a two-stage pipeline. First, we lift video pixels into a 4D human representation and then retarget to the humanoid morphology. Second, we propose GenMimic-a physics-aware reinforcement learning policy conditioned on 3D keypoints, and trained with symmetry regularization and keypoint-weighted tracking rewards. As a result, GenMimic can mimic human actions from noisy, generated videos. We curate GenMimicBench, a synthetic human-motion dataset generated using two video generation models across a spectrum of actions and contexts, establishing a benchmark for assessing zero-shot generalization and policy robustness. Extensive experiments demonstrate improvements over strong baselines in simulation and confirm coherent, physically stable motion tracking on a Unitree G1 humanoid robot without fine-tuning. This work offers a promising path to realizing the potential of video generation models as high-level policies for robot control.

Paperid: 717, https://arxiv.org/pdf/2512.04040.pdf

Abstract:
A truly interactive world model requires three key ingredients: real-time long-horizon streaming, consistent spatial memory, and precise user control. However, most existing approaches address only one of these aspects in isolation, as achieving all three simultaneously is highly challenging-for example, long-term memory mechanisms often degrade real-time performance. In this work, we present RELIC, a unified framework that tackles these three challenges altogether. Given a single image and a text description, RELIC enables memory-aware, long-duration exploration of arbitrary scenes in real time. Built upon recent autoregressive video-diffusion distillation techniques, our model represents long-horizon memory using highly compressed historical latent tokens encoded with both relative actions and absolute camera poses within the KV cache. This compact, camera-aware memory structure supports implicit 3D-consistent content retrieval and enforces long-term coherence with minimal computational overhead. In parallel, we fine-tune a bidirectional teacher video model to generate sequences beyond its original 5-second training horizon, and transform it into a causal student generator using a new memory-efficient self-forcing paradigm that enables full-context distillation over long-duration teacher as well as long student self-rollouts. Implemented as a 14B-parameter model and trained on a curated Unreal Engine-rendered dataset, RELIC achieves real-time generation at 16 FPS while demonstrating more accurate action following, more stable long-horizon streaming, and more robust spatial-memory retrieval compared with prior work. These capabilities establish RELIC as a strong foundation for the next generation of interactive world modeling.

Paperid: 718, https://arxiv.org/pdf/2512.01728.pdf

Abstract:
This paper investigates the detection of misinformation, which deceives readers by explicitly fabricating misleading content or implicitly omitting important information necessary for informed judgment. While the former has been extensively studied, omission-based deception remains largely overlooked, even though it can subtly guide readers toward false conclusions under the illusion of completeness. To pioneer in this direction, this paper presents OmiGraph, the first omission-aware framework for misinformation detection. Specifically, OmiGraph constructs an omission-aware graph for the target news by utilizing a contextual environment that captures complementary perspectives of the same event, thereby surfacing potentially omitted contents. Based on this graph, omission-oriented relation modeling is then proposed to identify the internal contextual dependencies, as well as the dynamic omission intents, formulating a comprehensive omission relation representation. Finally, to extract omission patterns for detection, OmiGraph introduces omission-aware message-passing and aggregation that establishes holistic deception perception by integrating the omission contents and relations. Experiments show that, by considering the omission perspective, our approach attains remarkable performance, achieving average improvements of +5.4% F1 and +5.3% ACC on two large-scale benchmarks.

Paperid: 719, https://arxiv.org/pdf/2512.01629.pdf

Abstract:
Articulated 3D objects are critical for embodied AI, robotics, and interactive scene understanding, yet creating simulation-ready assets remains labor-intensive and requires expert modeling of part hierarchies and motion structures. We introduce SPARK, a framework for reconstructing physically consistent, kinematic part-level articulated objects from a single RGB image. Given an input image, we first leverage VLMs to extract coarse URDF parameters and generate part-level reference images. We then integrate the part-image guidance and the inferred structure graph into a generative diffusion transformer to synthesize consistent part and complete shapes of articulated objects. To further refine the URDF parameters, we incorporate differentiable forward kinematics and differentiable rendering to optimize joint types, axes, and origins under VLM-generated open-state supervision. Extensive experiments show that SPARK produces high-quality, simulation-ready articulated assets across diverse categories, enabling downstream applications such as robotic manipulation and interaction modeling. Project page: https://heyumeng.com/SPARK/index.html.

Paperid: 720, https://arxiv.org/pdf/2512.00591.pdf

Abstract:
Hardware Trojans (HT s) are a persistent threat to integrated circuits, especially when inserted at the register-transfer level (RTL). Existing methods typically first convert the design into a graph, such as a gate-level netlist or an RTL-derived dataflow graph (DFG), and then use a graph neural network (GNN ) to obtain an embedding of that graph, which (i) loses compact RTL semantics, (ii) relies on shallow GNNs with limited receptive field, and (iii) is largely restricted to coarse, module-level binary HT detection. We propose TrojanLoC, an LLM-based framework for RTL-level HT localization. We use an RTL-finetuned LLM to derive module-level and line-level embeddings directly from RTL code, capturing both global design context and local semantics. Next, we train task-specific classifiers on these embeddings to perform module-level Trojan detection, type prediction, and fine-grained line-level localization. We also introduce TrojanInS, a large synthetic dataset of RTL designs with systematically injected Trojans from four effect-based categories, each accompanied by precise line-level annotations. Our experiments show that TrojanLoC achieves strong module-level performance, reaching 0.99 F1-score for Trojan detection, up to 0.68 higher than baseline, and 0.84 macro-F1 for Trojan-type classification. At the line level, TrojanLoc further achieves up to 0.93 macro-F1, enabling fine-grained localization of Trojan-relevant RTL lines

Paperid: 721, https://arxiv.org/pdf/2512.00119.pdf

Abstract:
Graph neural networks (GNNs) have shown promise in hardware security by learning structural motifs from netlist graphs. However, this reliance on motifs makes GNNs vulnerable to adversarial netlist rewrites; even small-scale edits can mislead GNN predictions. Existing adversarial approaches, ranging from synthesis-recipe perturbations to gate transformations, come with high design overheads. We present NetDeTox, an automated end-to-end framework that orchestrates large language models (LLMs) with reinforcement learning (RL) in a systematic manner, enabling focused local rewriting. The RL agent identifies netlist components critical for GNN-based reasoning, while the LLM devises rewriting plans to diversify motifs that preserve functionality. Iterative feedback between the RL and LLM stages refines adversarial rewritings to limit overheads. Compared to the SOTA work AttackGNN, NetDeTox successfully degrades the effectiveness of all security schemes with fewer rewrites and substantially lower area overheads (reductions of 54.50% for GNN-RE, 25.44% for GNN4IP, and 41.04% for OMLA, respectively). For GNN4IP, ours can even optimize/reduce the original benchmarks' area, in particular for larger circuits, demonstrating the practicality and scalability of NetDeTox.

Paperid: 722, https://arxiv.org/pdf/2512.07472.pdf

Abstract:
Vision-Language-Action (VLA) models have shown great performance in robotic manipulation by mapping visual observations and language instructions directly to actions. However, they remain brittle under distribution shifts: when test scenarios change, VLAs often reproduce memorized trajectories instead of adapting to the updated scene, which is a failure mode we refer to as the "Memory Trap". This limitation stems from the end-to-end design, which lacks explicit 3D spatial reasoning and prevents reliable identification of actionable regions in unfamiliar environments. To compensate for this missing spatial understanding, 3D Spatial Affordance Fields (SAFs) can provide a geometric representation that highlights where interactions are physically feasible, offering explicit cues about regions the robot should approach or avoid. We therefore introduce Affordance Field Intervention (AFI), a lightweight hybrid framework that uses SAFs as an on-demand plug-in to guide VLA behavior. Our system detects memory traps through proprioception, repositions the robot to recent high-affordance regions, and proposes affordance-driven waypoints that anchor VLA-generated actions. A SAF-based scorer then selects trajectories with the highest cumulative affordance. Extensive experiments demonstrate that our method achieves an average improvement of 23.5% across different VLA backbones ($π_{0}$ and $π_{0.5}$) under out-of-distribution scenarios on real-world robotic platforms, and 20.2% on the LIBERO-Pro benchmark, validating its effectiveness in enhancing VLA robustness to distribution shifts.

Paperid: 723, https://arxiv.org/pdf/2512.07081.pdf

Abstract:
Heart failure (HF) is one of the leading causes of rehospitalization among older adults in the United States. Although clinical notes contain rich, detailed patient information and make up a large portion of electronic health records (EHRs), they remain underutilized for HF readmission risk analysis. Traditional computational models for HF readmission often rely on expert-crafted rules, medical thesauri, and ontologies to interpret clinical notes, which are typically written under time pressure and may contain misspellings, abbreviations, and domain-specific jargon. We present ClinNoteAgents, an LLM-based multi-agent framework that transforms free-text clinical notes into (1) structured representations of clinical and social risk factors for association analysis and (2) clinician-style abstractions for HF 30-day readmission prediction. We evaluate ClinNoteAgents on 3,544 notes from 2,065 patients (readmission rate=35.16%), demonstrating strong performance in extracting risk factors from free-text, identifying key contributing factors, and predicting readmission risk. By reducing reliance on structured fields and minimizing manual annotation and model training, ClinNoteAgents provides a scalable and interpretable approach to note-based HF readmission risk modeling in data-limited healthcare systems.

Paperid: 724, https://arxiv.org/pdf/2512.06255.pdf

Abstract:
Existing fine-grained image retrieval (FGIR) methods learn discriminative embeddings by adopting semantically sparse one-hot labels derived from category names as supervision. While effective on seen classes, such supervision overlooks the rich semantics encoded in category names, hindering the modeling of comparability among cross-category details and, in turn, limiting generalization to unseen categories. To tackle this, we introduce LaFG, a Language-driven framework for Fine-Grained Retrieval that converts class names into attribute-level supervision using large language models (LLMs) and vision-language models (VLMs). Treating each name as a semantic anchor, LaFG prompts an LLM to generate detailed, attribute-oriented descriptions. To mitigate attribute omission in these descriptions, it leverages a frozen VLM to project them into a vision-aligned space, clustering them into a dataset-wide attribute vocabulary while harvesting complementary attributes from related categories. Leveraging this vocabulary, a global prompt template selects category-relevant attributes, which are aggregated into category-specific linguistic prototypes. These prototypes supervise the retrieval model to steer

Paperid: 725, https://arxiv.org/pdf/2512.05230.pdf

Abstract:
Reasoning from diverse observations is a fundamental capability for generalist robot policies to operate in a wide range of environments. Despite recent advancements, many large-scale robotic policies still remain sensitive to key sources of observational variation such as changes in camera perspective, lighting, and the presence of distractor objects. We posit that the limited generalizability of these models arises from the substantial diversity required to robustly cover these quasistatic axes, coupled with the current scarcity of large-scale robotic datasets that exhibit rich variation across them. In this work, we propose to systematically examine what robots need to generalize across these challenging axes by introducing two key auxiliary tasks, state similarity and invariance to observational perturbations, applied to both demonstration data and static visual data. We then show that via these auxiliary tasks, leveraging both more-expensive robotic demonstration data and less-expensive, visually rich synthetic images generated from non-physics-based simulation (for example, Unreal Engine) can lead to substantial increases in generalization to unseen camera viewpoints, lighting configurations, and distractor conditions. Our results demonstrate that co-training on this diverse data improves performance by 18 percent over existing generative augmentation methods. For more information and videos, please visit https://invariance-cotraining.github.io

Paperid: 726, https://arxiv.org/pdf/2512.04753.pdf

Abstract:
Knowledge editing aims to update specific facts in large language models (LLMs) without full retraining. Prior efforts sought to tune the knowledge layers of LLMs, proving effective for making selective edits. However, a significant gap exists between their performance in controlled, teacher-forcing evaluations and their real-world effectiveness in lifelong learning scenarios, which greatly limits their practical applicability. This work's empirical analysis reveals two recurring issues associated with this gap: (1) Most traditional methods lead the edited model to overfit to the new fact, thereby degrading pre-trained capabilities; (2) There is a critical absence of a knowledge consolidation stage, leaving new facts insufficiently integrated into LLMs' inference-time behavior under autoregressive generation, thereby leading to a mismatch between parametric knowledge and actual generation behavior. To this end, we propose Edit-then-Consolidate, a novel knowledge editing paradigm that aims to bridge the gap between theoretical knowledge editing methods and their real-world applicability. Specifically, (1) our framework mitigates overfitting via Targeted Proximal Supervised Fine-Tuning (TPSFT) that localizes the edit via a trust-region objective to limit policy drift; (2) Then, a consolidation stage using Group Relative Policy Optimization (GRPO) aligns the edited knowledge with CoT-based inference policy by optimizing trajectory-level behavior under comprehensive reward signals. Extensive experiments demonstrate our framework consistently improves editing reliability and generalization under real-world evaluations, while better preserving locality and pre-trained capabilities.

Paperid: 727, https://arxiv.org/pdf/2512.04248.pdf

Abstract:
We introduce MVRoom, a controllable novel view synthesis (NVS) pipeline for 3D indoor scenes that uses multi-view diffusion conditioned on a coarse 3D layout. MVRoom employs a two-stage design in which the 3D layout is used throughout to enforce multi-view consistency. The first stage employs novel representations to effectively bridge the 3D layout and consistent image-based condition signals for multi-view generation. The second stage performs image-conditioned multi-view generation, incorporating a layout-aware epipolar attention mechanism to enhance multi-view consistency during the diffusion process. Additionally, we introduce an iterative framework that generates 3D scenes with varying numbers of objects and scene complexities by recursively performing multi-view generation (MVRoom), supporting text-to-scene generation. Experimental results demonstrate that our approach achieves high-fidelity and controllable 3D scene generation for NVS, outperforming state-of-the-art baseline methods both quantitatively and qualitatively. Ablation studies further validate the effectiveness of key components within our generation pipeline.

Paperid: 728, https://arxiv.org/pdf/2512.03335.pdf

Abstract:
Design generation, in its essence, is a step-by-step process where designers progressively refine and enhance their work through careful modifications. Despite this fundamental characteristic, existing approaches mainly treat design synthesis as a single-step generation problem, significantly underestimating the inherent complexity of the creative process. To bridge this gap, we propose a novel problem setting called Step-by-Step Layered Design Generation, which tasks a machine learning model with generating a design that adheres to a sequence of instructions from a designer. Leveraging recent advancements in multi-modal LLMs, we propose SLEDGE: Step-by-step LayEred Design GEnerator to model each update to a design as an atomic, layered change over its previous state, while being grounded in the instruction. To complement our new problem setting, we introduce a new evaluation suite, including a dataset and a benchmark. Our exhaustive experimental analysis and comparison with state-of-the-art approaches tailored to our new setup demonstrate the efficacy of our approach. We hope our work will attract attention to this pragmatic and under-explored research area.

Paperid: 729, https://arxiv.org/pdf/2512.02807.pdf

Abstract:
Aligning Large Language Models (LLMs) with human preferences typically relies on external supervision, which faces critical limitations: human annotations are scarce and subjective, reward models are vulnerable to reward hacking, and self-evaluation methods suffer from prompt sensitivity and biases. In this work, we propose stable rank, an intrinsic, annotation-free quality signal derived from model representations. Stable rank measures the effective dimensionality of hidden states by computing the ratio of total variance to dominant-direction variance, capturing quality through how information distributes across representation dimensions. Empirically, stable rank achieves 84.04% accuracy on RewardBench and improves task accuracy by an average of 11.3 percentage points over greedy decoding via Best-of-N sampling. Leveraging this insight, we introduce Stable Rank Group Relative Policy Optimization (SR-GRPO), which uses stable rank as a reward signal for reinforcement learning. Without external supervision, SR-GRPO improves Qwen2.5-1.5B-Instruct by 10% on STEM and 19% on mathematical reasoning, outperforming both learned reward models and self-evaluation baselines. Our findings demonstrate that quality signals can be extracted from internal model geometry, offering a path toward scalable alignment without external supervision.

Paperid: 730, https://arxiv.org/pdf/2512.01715.pdf

Abstract:
Vision-Language-Action (VLA) models trained with flow matching have demonstrated impressive capabilities on robotic manipulation tasks. However, their performance often degrades under distribution shift and on complex multi-step tasks, suggesting that the learned representations may not robustly capture task-relevant semantics. We introduce DiG-Flow, a principled framework that enhances VLA robustness through geometric regularization. Our key insight is that the distributional discrepancy between observation and action embeddings provides a meaningful geometric signal: lower transport cost indicates compatible representations, while higher cost suggests potential misalignment. DiG-Flow computes a discrepancy measure between empirical distributions of observation and action embeddings, maps it to a modulation weight via a monotone function, and applies residual updates to the observation embeddings before flow matching. Crucially, this intervention operates at the representation level without modifying the flow matching path or target vector field. We provide theoretical guarantees showing that discrepancy-guided training provably decreases the training objective, and that guided inference refinement converges with contraction. Empirically, DiG-Flow integrates into existing VLA architectures with negligible overhead and consistently improves performance, with particularly pronounced gains on complex multi-step tasks and under limited training data.

Paperid: 731, https://arxiv.org/pdf/2512.01334.pdf

Abstract:
Text-guided image-to-video (TI2V) generation has recently achieved remarkable progress, particularly in maintaining subject consistency and temporal coherence. However, existing methods still struggle to adhere to fine-grained prompt semantics, especially when prompts entail substantial transformations of the input image (e.g., object addition, deletion, or modification), a shortcoming we term semantic negligence. In a pilot study, we find that applying a Gaussian blur to the input image improves semantic adherence. Analyzing attention maps, we observe clearer foreground-background separation. From an energy perspective, this corresponds to a lower-entropy cross-attention distribution. Motivated by this, we introduce AlignVid, a training-free framework with two components: (i) Attention Scaling Modulation (ASM), which directly reweights attention via lightweight Q or K scaling, and (ii) Guidance Scheduling (GS), which applies ASM selectively across transformer blocks and denoising steps to reduce visual quality degradation. This minimal intervention improves prompt adherence while limiting aesthetic degradation. In addition, we introduce OmitI2V to evaluate semantic negligence in TI2V generation, comprising 367 human-annotated samples that span addition, deletion, and modification scenarios. Extensive experiments demonstrate that AlignVid can enhance semantic fidelity.

Paperid: 732, https://arxiv.org/pdf/2512.00849.pdf

Abstract:
Clustering non-independent and identically distributed (non-IID) data under local differential privacy (LDP) in federated settings presents a critical challenge: preserving privacy while maintaining accuracy without iterative communication. Existing one-shot methods rely on unstable pairwise centroid distances or neighborhood rankings, degrading severely under strong LDP noise and data heterogeneity. We present Gravitational Federated Clustering (GFC), a novel approach to privacy-preserving federated clustering that overcomes the limitations of distance-based methods under varying LDP. Addressing the critical challenge of clustering non-IID data with diverse privacy guarantees, GFC transforms privatized client centroids into a global gravitational potential field where true cluster centers emerge as topologically persistent singularities. Our framework introduces two key innovations: (1) a client-side compactness-aware perturbation mechanism that encodes local cluster geometry as "mass" values, and (2) a server-side topological aggregation phase that extracts stable centroids through persistent homology analysis of the potential field's superlevel sets. Theoretically, we establish a closed-form bound between the privacy budget $ε$ and centroid estimation error, proving the potential field's Lipschitz smoothing properties exponentially suppress noise in high-density regions. Empirically, GFC outperforms state-of-the-art methods on ten benchmarks, especially under strong LDP constraints ($ε< 1$), while maintaining comparable performance at lower privacy budgets. By reformulating federated clustering as a topological persistence problem in a synthetic physics-inspired space, GFC achieves unprecedented privacy-accuracy trade-offs without iterative communication, providing a new perspective for privacy-preserving distributed learning.

Paperid: 733, https://arxiv.org/pdf/2512.06368.pdf

Abstract:
Monocular dynamic video reconstruction faces significant challenges in dynamic human scenes due to geometric inconsistencies and resolution degradation issues. Existing methods lack 3D human structural understanding, producing geometrically inconsistent results with distorted limb proportions and unnatural human-object fusion, while memory-constrained downsampling causes human boundary drift toward background geometry. To address these limitations, we propose to incorporate hybrid geometric priors that combine SMPL human body models with monocular depth estimation. Our approach leverages structured human priors to maintain surface consistency while capturing fine-grained geometric details in human regions. We introduce Human3R, featuring a hierarchical pipeline with refinement components that processes full-resolution images for overall scene geometry, then applies strategic cropping and cross-attention fusion for human-specific detail enhancement. The method integrates SMPL priors through a Feature Fusion Module to ensure geometrically plausible reconstruction while preserving fine-grained human boundaries. Extensive experiments on TUM Dynamics and GTA-IM datasets demonstrate superior performance in dynamic human reconstruction.

Paperid: 734, https://arxiv.org/pdf/2512.05564.pdf

Abstract:
Recent advances in video generation have shown remarkable potential for constructing world simulators. However, current models still struggle to produce physically consistent results, particularly when handling large-scale or complex dynamics. This limitation arises primarily because existing approaches respond isotropically to physical prompts and neglect the fine-grained alignment between generated content and localized physical cues. To address these challenges, we propose ProPhy, a Progressive Physical Alignment Framework that enables explicit physics-aware conditioning and anisotropic generation. ProPhy employs a two-stage Mixture-of-Physics-Experts (MoPE) mechanism for discriminative physical prior extraction, where Semantic Experts infer semantic-level physical principles from textual descriptions, and Refinement Experts capture token-level physical dynamics. This mechanism allows the model to learn fine-grained, physics-aware video representations that better reflect underlying physical laws. Furthermore, we introduce a physical alignment strategy that transfers the physical reasoning capabilities of vision-language models (VLMs) into the Refinement Experts, facilitating a more accurate representation of dynamic physical phenomena. Extensive experiments on physics-aware video generation benchmarks demonstrate that ProPhy produces more realistic, dynamic, and physically coherent results than existing state-of-the-art methods.

Paperid: 735, https://arxiv.org/pdf/2512.03360.pdf

Abstract:
Logical reasoning is a core challenge in natural language understanding and a fundamental capability of artificial intelligence, underpinning scientific discovery, mathematical theorem proving, and complex decision-making. Despite the remarkable progress of large language models (LLMs), most current approaches still rely on forward reasoning paradigms, generating step-by-step rationales from premises to conclusions. However, such methods often suffer from redundant inference paths, hallucinated steps, and semantic drift, resulting in inefficient and unreliable reasoning. In this paper, we propose a novel framework, Hypothesis-driven Backward Logical Reasoning (HBLR). The core idea is to integrate confidence-aware symbolic translation with hypothesis-driven backward reasoning. In the translation phase, only high-confidence spans are converted into logical form, such as First-Order Logic (FOL), while uncertain content remains in natural language. A translation reflection module further ensures semantic fidelity by evaluating symbolic outputs and reverting lossy ones back to text when necessary. In the reasoning phase, HBLR simulates human deductive thinking by assuming the conclusion is true and recursively verifying its premises. A reasoning reflection module further identifies and corrects flawed inference steps, enhancing logical coherence. Extensive experiments on five reasoning benchmarks demonstrate that HBLR consistently outperforms strong baselines in both accuracy and efficiency.

Paperid: 736, https://arxiv.org/pdf/2512.02073.pdf

Abstract:
Graph contrastive learning (GCL) aims to learn discriminative semantic invariance by contrasting different views of the same graph that share critical topological patterns. However, existing GCL approaches with structural augmentations often struggle to identify task-relevant topological structures, let alone adapt to the varying coarse-to-fine topological granularities required across different downstream tasks. To remedy this issue, we introduce Hierarchical Topological Granularity Graph Contrastive Learning (HTG-GCL), a novel framework that leverages transformations of the same graph to generate multi-scale ring-based cellular complexes, embodying the concept of topological granularity, thereby generating diverse topological views. Recognizing that a certain granularity may contain misleading semantics, we propose a multi-granularity decoupled contrast and apply a granularity-specific weighting mechanism based on uncertainty estimation. Comprehensive experiments on various benchmarks demonstrate the effectiveness of HTG-GCL, highlighting its superior performance in capturing meaningful graph representations through hierarchical topological information.

Paperid: 737, https://arxiv.org/pdf/2512.01481.pdf

Abstract:
Although prevailing camera-controlled video generation models can produce cinematic results, lifting them directly to the generation of 3D-consistent and high-fidelity time-synchronized multi-view videos remains challenging, which is a pivotal capability for taming 4D worlds. Some works resort to data augmentation or test-time optimization, but these strategies are constrained by limited model generalization and scalability issues. To this end, we propose ChronosObserver, a training-free method including World State Hyperspace to represent the spatiotemporal constraints of a 4D world scene, and Hyperspace Guided Sampling to synchronize the diffusion sampling trajectories of multiple views using the hyperspace. Experimental results demonstrate that our method achieves high-fidelity and 3D-consistent time-synchronized multi-view videos generation without training or fine-tuning for diffusion models.

Paperid: 738, https://arxiv.org/pdf/2512.00794.pdf

Abstract:
Recent advances in surface reconstruction for 3D Gaussian Splatting (3DGS) have enabled remarkable geometric accuracy. However, their performance degrades in photometrically ambiguous regions such as reflective and textureless surfaces, where unreliable cues disrupt photometric consistency and hinder accurate geometry estimation. Reflected light is often partially polarized in a manner that reveals surface orientation, making polarization an optic complement to photometric cues in resolving such ambiguities. Therefore, we propose PolarGS, an optics-aware extension of RGB-based 3DGS that leverages polarization as an optical prior to resolve photometric ambiguities and enhance reconstruction accuracy. Specifically, we introduce two complementary modules: a polarization-guided photometric correction strategy, which ensures photometric consistency by identifying reflective regions via the Degree of Linear Polarization (DoLP) and refining reflective Gaussians with Color Refinement Maps; and a polarization-enhanced Gaussian densification mechanism for textureless area geometry recovery, which integrates both Angle and Degree of Linear Polarization (A/DoLP) into a PatchMatch-based depth completion process. This enables the back-projection and fusion of new Gaussians, leading to more complete reconstruction. PolarGS is framework-agnostic and achieves superior geometric accuracy compared to state-of-the-art methods.

Paperid: 739, https://arxiv.org/pdf/2512.00098.pdf

Abstract:
This paper introduces GAMBiT (Guarding Against Malicious Biased Threats), a cognitive-informed cyber defense framework that leverages deviations from human rationality as a new defensive surface. Conventional cyber defenses assume rational, utility-maximizing attackers, yet real-world adversaries exhibit cognitive constraints and biases that shape their interactions with complex digital systems. GAMBiT embeds insights from cognitive science into cyber environments through cognitive triggers, which activate biases such as loss aversion, base-rate neglect, and sunk-cost fallacy, and through newly developed cognitive sensors that infer attackers' cognitive states from behavioral and network data. Three rounds of human-subject experiments (total n=61) in a simulated small business network demonstrate that these manipulations significantly disrupt attacker performance, reducing mission progress, diverting actions off the true attack path, and increasing detectability. These results demonstrate that cognitive biases can be systematically triggered to degrade the attacker's efficiency and enhance the defender's advantage. GAMBiT establishes a new paradigm in which the attacker's mind becomes part of the battlefield and cognitive manipulation becomes a proactive vector for cyber defense.

Paperid: 740, https://arxiv.org/pdf/2512.07461.pdf

Abstract:
We introduce Native Parallel Reasoner (NPR), a teacher-free framework that enables Large Language Models (LLMs) to self-evolve genuine parallel reasoning capabilities. NPR transforms the model from sequential emulation to native parallel cognition through three key innovations: 1) a self-distilled progressive training paradigm that transitions from ``cold-start'' format discovery to strict topological constraints without external supervision; 2) a novel Parallel-Aware Policy Optimization (PAPO) algorithm that optimizes branching policies directly within the execution graph, allowing the model to learn adaptive decomposition via trial and error; and 3) a robust NPR Engine that refactors memory management and flow control of SGLang to enable stable, large-scale parallel RL training. Across eight reasoning benchmarks, NPR trained on Qwen3-4B achieves performance gains of up to 24.5% and inference speedups up to 4.6x. Unlike prior baselines that often fall back to autoregressive decoding, NPR demonstrates 100% genuine parallel execution, establishing a new standard for self-evolving, efficient, and scalable agentic reasoning.

Paperid: 741, https://arxiv.org/pdf/2512.07433.pdf

Abstract:
Graph hyperdimensional computing (HDC) has emerged as a promising paradigm for cognitive tasks, emulating brain-like computation with high-dimensional vectors known as hypervectors. While HDC offers robustness and efficiency on graph-structured data, its fairness implications remain largely unexplored. In this paper, we study fairness in graph HDC, where biases in data representation and decision rules can lead to unequal treatment of different groups. We show how hypervector encoding and similarity-based classification can propagate or even amplify such biases, and we propose a fairness-aware training framework, FairGHDC, to mitigate them. FairGHDC introduces a bias correction term, derived from a gap-based demographic-parity regularizer, and converts it into a scalar fairness factor that scales the update of the class hypervector for the ground-truth label. This enables debiasing directly in the hypervector space without modifying the graph encoder or requiring backpropagation. Experimental results on six benchmark datasets demonstrate that FairGHDC substantially reduces demographic-parity and equal-opportunity gaps while maintaining accuracy comparable to standard GNNs and fairness-aware GNNs. At the same time, FairGHDC preserves the computational advantages of HDC, achieving up to about one order of magnitude ($\approx 10\times$) speedup in training time on GPU compared to GNN and fairness-aware GNN baselines.

Paperid: 742, https://arxiv.org/pdf/2512.07375.pdf

Abstract:
Large language models (LLMs) possess vast knowledge acquired from extensive training corpora, but they often cannot remove specific pieces of information when needed, which makes it hard to handle privacy, bias mitigation, and knowledge correction. Traditional model unlearning approaches require computationally expensive fine-tuning or direct weight editing, making them impractical for real-world deployment. In this work, we introduce LoRA-based Unlearning with Negative Examples (LUNE), a lightweight framework that performs negative-only unlearning by updating only low-rank adapters while freezing the backbone, thereby localizing edits and avoiding disruptive global changes. Leveraging Low-Rank Adaptation (LoRA), LUNE targets intermediate representations to suppress (or replace) requested knowledge with an order-of-magnitude lower compute and memory than full fine-tuning or direct weight editing. Extensive experiments on multiple factual unlearning tasks show that LUNE: (I) achieves effectiveness comparable to full fine-tuning and memory-editing methods, and (II) reduces computational cost by about an order of magnitude.

Paperid: 743, https://arxiv.org/pdf/2512.07374.pdf

Abstract:
Unlearning in large foundation models (e.g., LLMs) is essential for enabling dynamic knowledge updates, enforcing data deletion rights, and correcting model behavior. However, existing unlearning methods often require full-model fine-tuning or access to the original training data, which limits their scalability and practicality. In this work, we introduce Recover-to-Forget (R2F), a novel framework for efficient unlearning in LLMs based on reconstructing full-model gradient directions from low-rank LoRA adapter updates. Rather than performing backpropagation through the full model, we compute gradients with respect to LoRA parameters using multiple paraphrased prompts and train a gradient decoder to approximate the corresponding full-model gradients. To ensure applicability to larger or black-box models, the decoder is trained on a proxy model and transferred to target models. We provide a theoretical analysis of cross-model generalization and demonstrate that our method achieves effective unlearning while preserving general model performance. Experimental results demonstrate that R2F offers a scalable and lightweight alternative for unlearning in pretrained LLMs without requiring full retraining or access to internal parameters.

Paperid: 744, https://arxiv.org/pdf/2512.07289.pdf

Abstract:
In addressing the challenge of Crystal Structure Prediction (CSP), symmetry-aware deep learning models, particularly diffusion models, have been extensively studied, which treat CSP as a conditional generation task. However, ensuring permutation, rotation, and periodic translation equivariance during diffusion process remains incompletely addressed. In this work, we propose EquiCSP, a novel equivariant diffusion-based generative model. We not only address the overlooked issue of lattice permutation equivariance in existing models, but also develop a unique noising algorithm that rigorously maintains periodic translation equivariance throughout both training and inference processes. Our experiments indicate that EquiCSP significantly surpasses existing models in terms of generating accurate structures and demonstrates faster convergence during the training process.

Paperid: 745, https://arxiv.org/pdf/2512.07184.pdf

Abstract:
As multimodal data proliferates across diverse real-world applications, leveraging heterogeneous information such as texts and timestamps for accurate time series forecasting (TSF) has become a critical challenge. While diffusion models demonstrate exceptional performance in generation tasks, their application to TSF remains largely confined to modeling single-modality numerical sequences, overlooking the abundant cross-modal signals inherent in complex heterogeneous data. To address this gap, we propose UniDiff, a unified diffusion framework for multimodal time series forecasting. To process the numerical sequence, our framework first tokenizes the time series into patches, preserving local temporal dynamics by mapping each patch to an embedding space via a lightweight MLP. At its core lies a unified and parallel fusion module, where a single cross-attention mechanism adaptively weighs and integrates structural information from timestamps and semantic context from texts in one step, enabling a flexible and efficient interplay between modalities. Furthermore, we introduce a novel classifier-free guidance mechanism designed for multi-source conditioning, allowing for decoupled control over the guidance strength of textual and temporal information during inference, which significantly enhances model robustness. Extensive experiments on real-world benchmark datasets across eight domains demonstrate that the proposed UniDiff model achieves state-of-the-art performance.

Paperid: 746, https://arxiv.org/pdf/2512.06793.pdf

Abstract:
Real-time stereo matching methods primarily focus on enhancing in-domain performance but often overlook the critical importance of generalization in real-world applications. In contrast, recent stereo foundation models leverage monocular foundation models (MFMs) to improve generalization, but typically suffer from substantial inference latency. To address this trade-off, we propose Generalized Geometry Encoding Volume (GGEV), a novel real-time stereo matching network that achieves strong generalization. We first extract depth-aware features that encode domain-invariant structural priors as guidance for cost aggregation. Subsequently, we introduce a Depth-aware Dynamic Cost Aggregation (DDCA) module that adaptively incorporates these priors into each disparity hypothesis, effectively enhancing fragile matching relationships in unseen scenes. Both steps are lightweight and complementary, leading to the construction of a generalized geometry encoding volume with strong generalization capability. Experimental results demonstrate that our GGEV surpasses all existing real-time methods in zero-shot generalization capability, and achieves state-of-the-art performance on the KITTI 2012, KITTI 2015, and ETH3D benchmarks.

Paperid: 747, https://arxiv.org/pdf/2512.06684.pdf

Abstract:
Volume electron microscopy (vEM) enables nanoscale 3D imaging of biological structures but remains constrained by acquisition trade-offs, leading to anisotropic volumes with limited axial resolution. Existing deep learning methods seek to restore isotropy by leveraging lateral priors, yet their assumptions break down for morphologically anisotropic structures. We present EMGauss, a general framework for 3D reconstruction from planar scanned 2D slices with applications in vEM, which circumvents the inherent limitations of isotropy-based approaches. Our key innovation is to reframe slice-to-3D reconstruction as a 3D dynamic scene rendering problem based on Gaussian splatting, where the progression of axial slices is modeled as the temporal evolution of 2D Gaussian point clouds. To enhance fidelity in data-sparse regimes, we incorporate a Teacher-Student bootstrapping mechanism that uses high-confidence predictions on unobserved slices as pseudo-supervisory signals. Compared with diffusion- and GAN-based reconstruction methods, EMGauss substantially improves interpolation quality, enables continuous slice synthesis, and eliminates the need for large-scale pretraining. Beyond vEM, it potentially provides a generalizable slice-to-3D solution across diverse imaging domains.

Paperid: 748, https://arxiv.org/pdf/2512.06533.pdf

Abstract:
Decoding-based regression, which reformulates regression as a sequence generation task, has emerged as a promising paradigm of applying large language models for numerical prediction. However, its progress is hindered by the misalignment between discrete token-level objectives (e.g., cross-entropy) and continuous numerical values. Existing approaches relying on token-level constraints often fail to capture the global magnitude of the target value, limiting their precision and generalization. In this paper, we propose to unlock the potential of decoding-based regression via Reinforcement Learning (RL). We formulate the generation process as a Markov Decision Process, utilizing sequence-level rewards to enforce global numerical coherence. Extensive experiments on tabular regression and code metric regression demonstrate that our method (specifically with ReMax and GRPO) consistently outperforms both state-of-the-art token-level baselines and traditional regression heads, showing the superiority of introducing sequence-level signals. Our analysis further reveals that RL significantly enhances sampling efficiency and predictive precision, establishing decoding-based regression as a robust and accurate paradigm for general-purpose numerical prediction.

Paperid: 749, https://arxiv.org/pdf/2512.04857.pdf

Abstract:
Autoregressive (AR) visual generation has emerged as a powerful paradigm for image and multimodal synthesis, owing to its scalability and generality. However, existing AR image generation suffers from severe memory bottlenecks due to the need to cache all previously generated visual tokens during decoding, leading to both high storage requirements and low throughput. In this paper, we introduce \textbf{LineAR}, a novel, training-free progressive key-value (KV) cache compression pipeline for autoregressive image generation. By fully exploiting the intrinsic characteristics of visual attention, LineAR manages the cache at the line level using a 2D view, preserving the visual dependency regions while progressively evicting less-informative tokens that are harmless for subsequent line generation, guided by inter-line attention. LineAR enables efficient autoregressive (AR) image generation by utilizing only a few lines of cache, achieving both memory savings and throughput speedup, while maintaining or even improving generation quality. Extensive experiments across six autoregressive image generation models, including class-conditional and text-to-image generation, validate its effectiveness and generality. LineAR improves ImageNet FID from 2.77 to 2.68 and COCO FID from 23.85 to 22.86 on LlamaGen-XL and Janus-Pro-1B, while retaining only 1/6 KV cache. It also improves DPG on Lumina-mGPT-768 with just 1/8 KV cache. Additionally, LineAR achieves significant memory and throughput gains, including up to 67.61% memory reduction and 7.57x speedup on LlamaGen-XL, and 39.66% memory reduction and 5.62x speedup on Janus-Pro-7B.

Paperid: 750, https://arxiv.org/pdf/2512.04380.pdf

Abstract:
The use of terahertz (THz) communications with massive multiple input multiple output (MIMO) systems in 6G can potentially provide high data rates and low latency communications. However, accurate channel estimation in THz frequencies presents significant challenges due to factors such as high propagation losses, sensitivity to environmental obstructions, and strong atmospheric absorption. These challenges are particularly pronounced in urban environments, where traditional channel estimation methods often fail to deliver reliable results, particularly in complex non-line-of-sight (NLoS) scenarios. This paper introduces a novel vision-based channel estimation technique that integrates causal reasoning into urban THz communication systems. The proposed method combines computer vision algorithms with variational causal dynamics (VCD) to analyze real-time images of the urban environment, allowing for a deeper understanding of the physical factors that influence THz signal propagation. By capturing the complex, dynamic interactions between physical objects (such as buildings, trees, and vehicles) and the transmitted signals, the model can predict the channel with up to twice the accuracy of conventional methods. This model improves estimation accuracy and demonstrates superior generalization performance. Hence, it can provide reliable predictions even in previously unseen urban environments. The effectiveness of the proposed method is particularly evident in NLoS conditions, where it significantly outperforms traditional methods such as by accounting for indirect signal paths, such as reflections and diffractions. Simulation results confirm that the proposed vision-based approach surpasses conventional artificial intelligence (AI)-based estimation techniques in accuracy and robustness, showing a substantial improvement across various dynamic urban scenarios.

Paperid: 751, https://arxiv.org/pdf/2512.03318.pdf

Authors:Chandler Smith, Marwa Abdulhai, Manfred Diaz, Marko Tesic, Rakshit S. Trivedi, Alexander Sasha Vezhnevets, Lewis Hammond, Jesse Clifton, Minsuk Chang, Edgar A. Duéñez-Guzmán, John P. Agapiou, Jayd Matyas, Danny Karmon, Akash Kundu, Aliaksei Korshuk, Ananya Ananya, Arrasy Rahman, Avinaash Anand Kulandaivel, Bain McHale, Beining Zhang, Buyantuev Alexander, Carlos Saith Rodriguez Rojas, Caroline Wang, Chetan Talele, Chenao Liu, Chichen Lin, Diana Riazi, Di Yang Shi, Emanuel Tewolde, Elizaveta Tennant, Fangwei Zhong, Fuyang Cui, Gang Zhao, Gema Parreño Piqueras, Hyeonggeun Yun, Ilya Makarov, Jiaxun Cui, Jebish Purbey, Jim Dilkes, Jord Nguyen, Lingyun Xiao, Luis Felipe Giraldo, Manuela Chacon-Chamorro, Manuel Sebastian Rios Beltran, Marta Emili García Segura, Mengmeng Wang, Mogtaba Alim, Nicanor Quijano, Nico Schiavone, Olivia Macmillan-Scott, Oswaldo Peña, Peter Stone, Ram Mohan Rao Kadiyala, Rolando Fernandez, Ruben Manrique, Sunjia Lu, Sheila A. McIlraith, Shamika Dhuri, Shuqing Shi, Siddhant Gupta, Sneheel Sarangi, Sriram Ganapathi Subramanian, Taehun Cha, Toryn Q. Klassen, Wenming Tu, Weijian Fan, Wu Ruiyang, Xue Feng, Yali Du, Yang Liu, Yiding Wang, Yipeng Kang, Yoonchang Sung, Yuxuan Chen, Zhaowei Zhang, Zhihan Wang, Zhiqiang Wu, Ziang Chen, Zilong Zheng, Zixia Jia, Ziyan Wang, Dylan Hadfield-Menell, Natasha Jaques, Tim Baarslag, Jose Hernandez-Orallo, Joel Z. Leibo

Abstract:
Large Language Model (LLM) agents have demonstrated impressive capabilities for social interaction and are increasingly being deployed in situations where they might engage with both human and artificial agents. These interactions represent a critical frontier for LLM-based agents, yet existing evaluation methods fail to measure how well these capabilities generalize to novel social situations. In this paper, we introduce a method for evaluating the ability of LLM-based agents to cooperate in zero-shot, mixed-motive environments using Concordia, a natural language multi-agent simulation environment. Our method measures general cooperative intelligence by testing an agent's ability to identify and exploit opportunities for mutual gain across diverse partners and contexts. We present empirical results from the NeurIPS 2024 Concordia Contest, where agents were evaluated on their ability to achieve mutual gains across a suite of diverse scenarios ranging from negotiation to collective action problems. Our findings reveal significant gaps between current agent capabilities and the robust generalization required for reliable cooperation, particularly in scenarios demanding persuasion and norm enforcement.

Paperid: 752, https://arxiv.org/pdf/2512.03052.pdf

Abstract:
We present LATTICE, a new framework for high-fidelity 3D asset generation that bridges the quality and scalability gap between 3D and 2D generative models. While 2D image synthesis benefits from fixed spatial grids and well-established transformer architectures, 3D generation remains fundamentally more challenging due to the need to predict both spatial structure and detailed geometric surfaces from scratch. These challenges are exacerbated by the computational complexity of existing 3D representations and the lack of structured and scalable 3D asset encoding schemes. To address this, we propose VoxSet, a semi-structured representation that compresses 3D assets into a compact set of latent vectors anchored to a coarse voxel grid, enabling efficient and position-aware generation. VoxSet retains the simplicity and compression advantages of prior VecSet methods while introducing explicit structure into the latent space, allowing positional embeddings to guide generation and enabling strong token-level test-time scaling. Built upon this representation, LATTICE adopts a two-stage pipeline: first generating a sparse voxelized geometry anchor, then producing detailed geometry using a rectified flow transformer. Our method is simple at its core, but supports arbitrary resolution decoding, low-cost training, and flexible inference schemes, achieving state-of-the-art performance on various aspects, and offering a significant step toward scalable, high-quality 3D asset creation.

Paperid: 753, https://arxiv.org/pdf/2512.03004.pdf

Abstract:
Autonomous driving needs fast, scalable 4D reconstruction and re-simulation for training and evaluation, yet most methods for dynamic driving scenes still rely on per-scene optimization, known camera calibration, or short frame windows, making them slow and impractical. We revisit this problem from a feedforward perspective and introduce \textbf{Driving Gaussian Grounded Transformer (DGGT)}, a unified framework for pose-free dynamic scene reconstruction. We note that the existing formulations, treating camera pose as a required input, limit flexibility and scalability. Instead, we reformulate pose as an output of the model, enabling reconstruction directly from sparse, unposed images and supporting an arbitrary number of views for long sequences. Our approach jointly predicts per-frame 3D Gaussian maps and camera parameters, disentangles dynamics with a lightweight dynamic head, and preserves temporal consistency with a lifespan head that modulates visibility over time. A diffusion-based rendering refinement further reduces motion/interpolation artifacts and improves novel-view quality under sparse inputs. The result is a single-pass, pose-free algorithm that achieves state-of-the-art performance and speed. Trained and evaluated on large-scale driving benchmarks (Waymo, nuScenes, Argoverse2), our method outperforms prior work both when trained on each dataset and in zero-shot transfer across datasets, and it scales well as the number of input frames increases.

Paperid: 754, https://arxiv.org/pdf/2512.02369.pdf

Abstract:
Domain generalization for semantic segmentation aims to mitigate the degradation in model performance caused by domain shifts. However, in many real-world scenarios, we are unable to access the model parameters and architectural details due to privacy concerns and security constraints. Traditional fine-tuning or adaptation is hindered, leading to the demand for input-level strategies that can enhance generalization without modifying model weights. To this end, we propose a \textbf{S}tyle-\textbf{A}daptive \textbf{GE}neralization framework (\textbf{SAGE}), which improves the generalization of frozen models under privacy constraints. SAGE learns to synthesize visual prompts that implicitly align feature distributions across styles instead of directly fine-tuning the backbone. Specifically, we first utilize style transfer to construct a diverse style representation of the source domain, thereby learning a set of style characteristics that can cover a wide range of visual features. Then, the model adaptively fuses these style cues according to the visual context of each input, forming a dynamic prompt that harmonizes the image appearance without touching the interior of the model. Through this closed-loop design, SAGE effectively bridges the gap between frozen model invariance and the diversity of unseen domains. Extensive experiments on five benchmark datasets demonstrate that SAGE achieves competitive or superior performance compared to state-of-the-art methods under privacy constraints and outperforms full fine-tuning baselines in all settings.

Paperid: 755, https://arxiv.org/pdf/2512.02350.pdf

Abstract:
Offline Federated Reinforcement Learning (FRL), a marriage of federated learning and offline reinforcement learning, has attracted increasing interest recently. Albeit with some advancement, we find that the performance of most existing offline FRL methods drops dramatically when provided with mixed-quality data, that is, the logging behaviors (offline data) are collected by policies with varying qualities across clients. To overcome this limitation, this paper introduces a new vote-based offline FRL framework, named FOVA. It exploits a \emph{vote mechanism} to identify high-return actions during local policy evaluation, alleviating the negative effect of low-quality behaviors from diverse local learning policies. Besides, building on advantage-weighted regression (AWR), we construct consistent local and global training objectives, significantly enhancing the efficiency and stability of FOVA. Further, we conduct an extensive theoretical analysis and rigorously show that the policy learned by FOVA enjoys strict policy improvement over the behavioral policy. Extensive experiments corroborate the significant performance gains of our proposed algorithm over existing baselines on widely used benchmarks.

Paperid: 756, https://arxiv.org/pdf/2512.01426.pdf

Abstract:
Leveraging pre-trained Diffusion Transformers (DiTs) for high-resolution (HR) image synthesis often leads to spatial layout collapse and degraded texture fidelity. Prior work mitigates these issues with complex pipelines that first perform a base-resolution (i.e., training-resolution) denoising process to guide HR generation. We instead explore the intrinsic generative mechanisms of DiTs and propose ResDiT, a training-free method that scales resolution efficiently. We identify the core factor governing spatial layout, position embeddings (PEs), and show that the original PEs encode incorrect positional information when extrapolated to HR, which triggers layout collapse. To address this, we introduce a PE scaling technique that rectifies positional encoding under resolution changes. To further remedy low-fidelity details, we develop a local-enhancement mechanism grounded in base-resolution local attention. We design a patch-level fusion module that aggregates global and local cues, together with a Gaussian-weighted splicing strategy that eliminates grid artifacts. Comprehensive evaluations demonstrate that ResDiT consistently delivers high-fidelity, high-resolution image synthesis and integrates seamlessly with downstream tasks, including spatially controlled generation.

Paperid: 757, https://arxiv.org/pdf/2512.00968.pdf

Abstract:
Ranking relevance is a fundamental task in search engines, aiming to identify the items most relevant to a given user query. Traditional relevance models typically produce scalar scores or directly predict relevance labels, limiting both interpretability and the modeling of complex relevance signals. Inspired by recent advances in Chain-of-Thought (CoT) reasoning for complex tasks, we investigate whether explicit reasoning can enhance both interpretability and performance in relevance modeling. However, existing reasoning-based Generative Relevance Models (GRMs) primarily rely on supervised fine-tuning on large amounts of human-annotated or synthetic CoT data, which often leads to limited generalization. Moreover, domain-agnostic, free-form reasoning tends to be overly generic and insufficiently grounded, limiting its potential to handle the diverse and ambiguous cases prevalent in open-domain search. In this work, we formulate relevance modeling in Xiaohongshu search as a reasoning task and introduce a Reinforcement Learning (RL)-based training framework to enhance the grounded reasoning capabilities of GRMs. Specifically, we incorporate practical business-specific relevance criteria into the multi-step reasoning prompt design and propose Stepwise Advantage Masking (SAM), a lightweight process-supervision strategy which facilitates effective learning of these criteria through improved credit assignment. To enable industrial deployment, we further distill the large-scale RL-tuned model to a lightweight version suitable for real-world search systems. Extensive experiments on industrial datasets, along with online A/B tests, demonstrate the effectiveness of our approach.

Paperid: 758, https://arxiv.org/pdf/2512.00775.pdf

Abstract:
Linear Temporal Logic (LTL) provides a rigorous framework for complex robotic tasks, yet existing methods often rely on accurate dynamics models or expensive online interactions. In this work, we address LTL-constrained control in a challenging offline, model-free setting, utilizing only fixed, task-agnostic datasets of fragmented trajectories. We propose SAGAS, a novel framework combining graph-assisted trajectory stitching with automata-guided planning. First, we construct a latent reachability graph from a learned temporal-distance representation. To bridge the semantic gap, we augment this graph with certified anchor nodes and probabilistic soft labels. We then translate the specification into a Büchi automaton and search the implicit product space to derive a cost-minimal prefix-suffix plan. Finally, a subgoal-conditioned low-level policy is deployed to execute these latent waypoints. Experiments on OGBench locomotion domains demonstrate that SAGAS successfully synthesizes efficient trajectories for diverse LTL tasks, effectively bridging the gap between fragmented offline data and complex logical constraints.

Paperid: 759, https://arxiv.org/pdf/2512.00375.pdf

Abstract:
Existing motion planning methods often struggle with rapid-motion obstacles due to an insufficient understanding of environmental changes. To address this limitation, we propose integrating motion planners with Doppler LiDARs which provide not only ranging measurements but also instantaneous point velocities. However, this integration is nontrivial due to the dual requirements of high accuracy and high frequency. To this end, we introduce Doppler Planning Network (DPNet), which tracks and reacts to rapid obstacles using Doppler model-based learning. Particularly, we first propose a Doppler Kalman neural network (D-KalmanNet) to track the future states of obstacles under partially observable Gaussian state space model. We then leverage the estimated motions to construct a Doppler-tuned model predictive control (DT-MPC) framework for ego-motion planning, enabling runtime auto-tuning of the controller parameters. These two model-based learners allow DPNet to maintain lightweight while learning fast environmental changes using minimum data, and achieve both high frequency and high accuracy in tracking and planning. Experiments on both high-fidelity simulator and real-world datasets demonstrate the superiority of DPNet over extensive benchmark schemes.

Paperid: 760, https://arxiv.org/pdf/2512.00341.pdf

Abstract:
Evolutionary Algorithms (EAs) are widely used general-purpose optimization methods due to their domain independence. However, under a limited number of function evaluations (#FEs), the performance of EAs is quite sensitive to the quality of the initial population. Obtaining a high-quality initial population without problem-specific knowledge remains a significant challenge. To address this, this work proposes a general-purpose population initialization method, named mixture-of-experience for population initialization (MPI), for binary optimization problems where decision variables take values of 0 or 1. MPI leverages solving experiences from previously solved problems to generate high-quality initial populations for new problems using only a small number of FEs. Its main novelty lies in a general-purpose approach for representing, selecting, and transferring solving experiences without requiring problem-specific knowledge. Extensive experiments are conducted across six binary optimization problem classes, comprising three classic classes and three complex classes from real-world applications. The experience repository is constructed solely based on instances from the three classic classes, while the performance evaluation is performed across all six classes. The results demonstrate that MPI effectively transfers solving experiences to unseen problem classes (i.e., the complex ones) and higher-dimensional problem instances, significantly outperforming existing general-purpose population initialization methods.

Paperid: 761, https://arxiv.org/pdf/2512.07783.pdf

Abstract:
Recent reinforcement learning (RL) techniques have yielded impressive reasoning improvements in language models, yet it remains unclear whether post-training truly extends a model's reasoning ability beyond what it acquires during pre-training. A central challenge is the lack of control in modern training pipelines: large-scale pre-training corpora are opaque, mid-training is often underexamined, and RL objectives interact with unknown prior knowledge in complex ways. To resolve this ambiguity, we develop a fully controlled experimental framework that isolates the causal contributions of pre-training, mid-training, and RL-based post-training. Our approach employs synthetic reasoning tasks with explicit atomic operations, parseable step-by-step reasoning traces, and systematic manipulation of training distributions. We evaluate models along two axes: extrapolative generalization to more complex compositions and contextual generalization across surface contexts. Using this framework, we reconcile competing views on RL's effectiveness. We show that: 1) RL produces true capability gains (pass@128) only when pre-training leaves sufficient headroom and when RL data target the model's edge of competence, tasks at the boundary that are difficult but not yet out of reach. 2) Contextual generalization requires minimal yet sufficient pre-training exposure, after which RL can reliably transfer. 3) Mid-training significantly enhances performance under fixed compute compared with RL only, demonstrating its central but underexplored role in training pipelines. 4) Process-level rewards reduce reward hacking and improve reasoning fidelity. Together, these results clarify the interplay between pre-training, mid-training, and RL, offering a foundation for understanding and improving reasoning LM training strategies.

Paperid: 762, https://arxiv.org/pdf/2512.06846.pdf

Abstract:
Traditional approaches for smart contract analysis often rely on intermediate representations such as abstract syntax trees, control-flow graphs, or static single assignment form. However, these methods face limitations in capturing both semantic structures and control logic. Knowledge graphs, by contrast, offer a structured representation of entities and relations, enabling richer intermediate abstractions of contract code and supporting the use of graph query languages to identify rule-violating elements. This paper presents CKG-LLM, a framework for detecting access-control vulnerabilities in smart contracts. Leveraging the reasoning and code generation capabilities of large language models, CKG-LLM translates natural-language vulnerability patterns into executable queries over contract knowledge graphs to automatically locate vulnerable code elements. Experimental evaluation demonstrates that CKG-LLM achieves superior performance in detecting access-control vulnerabilities compared to existing tools. Finally, we discuss potential extensions of CKG-LLM as part of future research directions.

Paperid: 763, https://arxiv.org/pdf/2512.06688.pdf

Abstract:
Personalization is one of the next milestones in advancing AI capability and alignment. We introduce PersonaMem-v2, the state-of-the-art dataset for LLM personalization that simulates 1,000 realistic user-chatbot interactions on 300+ scenarios, 20,000+ user preferences, and 128k-token context windows, where most user preferences are implicitly revealed to reflect real-world interactions. Using this data, we investigate how reinforcement fine-tuning enables a model to improve its long-context reasoning capabilities for user understanding and personalization. We also develop a framework for training an agentic memory system, which maintains a single, human-readable memory that grows with each user over time. In our experiments, frontier LLMs still struggle with implicit personalization, achieving only 37-48% accuracy. While they support long context windows, reasoning remains the bottleneck for implicit personalization tasks. Using reinforcement fine-tuning, we successfully train Qwen3-4B to outperforms GPT-5, reaching 53% accuracy in implicit personalization. Moreover, our agentic memory framework achieves state-of-the-art 55% accuracy while using 16x fewer input tokens, relying on a 2k-token memory instead of full 32k conversation histories. These results underscore the impact of our dataset and demonstrate agentic memory as a scalable path toward real-world personalized intelligence.

Paperid: 764, https://arxiv.org/pdf/2512.06662.pdf

Abstract:
People can view the same image differently: they focus on different regions, objects, and details in varying orders and describe them in distinct linguistic styles. This leads to substantial variability in image descriptions. However, existing models for personalized image description focus on linguistic style alone, with no prior work leveraging individual viewing patterns. We address this gap by explicitly modeling personalized viewing behavior as a core factor in description generation. Our method, DEPER (DEscription-PERception persona encoder), learns a subject embedding that captures both linguistic style and viewing behavior, guided by an auxiliary attention-prediction task. A lightweight adapter aligns these embeddings with a frozen vision-language model, enabling few-shot personalization without retraining. Across four datasets spanning diverse viewing tasks and both short and detailed descriptions, DEPER achieves a 24% average improvement, showing that modeling personalized attention produces more human-aligned and high-quality descriptions. We posit that understanding how people see helps predict what they say; modeling human diversity in perception can improve both performance and human alignment in multimodal systems.

Paperid: 765, https://arxiv.org/pdf/2512.06179.pdf

Abstract:
Attached shadows occur on the surface of the occluder where light cannot reach because of self-occlusion. They are crucial for defining the three-dimensional structure of objects and enhancing scene understanding. Yet existing shadow detection methods mainly target cast shadows, and there are no dedicated datasets or models for detecting attached shadows. To address this gap, we introduce a framework that jointly detects cast and attached shadows by reasoning about their mutual relationship with scene illumination and geometry. Our system consists of a shadow detection module that predicts both shadow types separately, and a light estimation module that infers the light direction from the detected shadows. The estimated light direction, combined with surface normals, allows us to derive a geometry-consistent partial map that identifies regions likely to be self-occluded. This partial map is then fed back to refine shadow predictions, forming a closed-loop reasoning process that iteratively improves both shadow segmentation and light estimation. In order to train our method, we have constructed a dataset of 1,458 images with separate annotations for cast and attached shadows, enabling training and quantitative evaluation of both. Experimental results demonstrate that this iterative geometry-illumination reasoning substantially improves the detection of attached shadows, with at least 33% BER reduction, while maintaining strong full and cast shadow performance.

Paperid: 766, https://arxiv.org/pdf/2512.06174.pdf

Abstract:
Shadow generation aims to produce photorealistic shadows that are visually consistent with object geometry and scene illumination. In the physics of shadow formation, the occluder blocks some light rays casting from the light source that would otherwise arrive at the surface, creating a shadow that follows the silhouette of the occluder. However, such explicit physical modeling has rarely been used in deep-learning-based shadow generation. In this paper, we propose a novel framework that embeds explicit physical modeling - geometry and illumination - into deep-learning-based shadow generation. First, given a monocular RGB image, we obtain approximate 3D geometry in the form of dense point maps and predict a single dominant light direction. These signals allow us to recover fairly accurate shadow location and shape based on the physics of shadow formation. We then integrate this physics-based initial estimate into a diffusion framework that refines the shadow into a realistic, high-fidelity appearance while ensuring consistency with scene geometry and illumination. Trained on DESOBAV2, our model produces shadows that are both visually realistic and physically coherent, outperforming existing approaches, especially in scenes with complex geometry or ambiguous lighting.

Paperid: 767, https://arxiv.org/pdf/2512.04111.pdf

Abstract:
LLM-powered coding agents are reshaping the development paradigm. However, existing evaluation systems, neither traditional tests for humans nor benchmarks for LLMs, fail to capture this shift. They remain focused on well-defined algorithmic problems, which excludes problems where success depends on human-AI collaboration. Such collaborative problems not only require human reasoning to interpret complex contexts and guide solution strategies, but also demand AI efficiency for implementation. To bridge this gap, we introduce HAI-Eval, a unified benchmark designed to measure the synergy of human-AI partnership in coding. HAI-Eval's core innovation is its "Collaboration-Necessary" problem templates, which are intractable for both standalone LLMs and unaided humans, but solvable through effective collaboration. Specifically, HAI-Eval uses 45 templates to dynamically create tasks. It also provides a standardized IDE for human participants and a reproducible toolkit with 450 task instances for LLMs, ensuring an ecologically valid evaluation. We conduct a within-subject study with 45 participants and benchmark their performance against 5 state-of-the-art LLMs under 4 different levels of human intervention. Results show that standalone LLMs and unaided participants achieve poor pass rates (0.67% and 18.89%), human-AI collaboration significantly improves performance to 31.11%. Our analysis reveals an emerging co-reasoning partnership. This finding challenges the traditional human-tool hierarchy by showing that strategic breakthroughs can originate from either humans or AI. HAI-Eval establishes not only a challenging benchmark for next-generation coding agents but also a grounded, scalable framework for assessing core developer competencies in the AI era. Our benchmark and interactive demo will be openly accessible.

Paperid: 768, https://arxiv.org/pdf/2512.03575.pdf

Abstract:
Distinct from attention-based compression methods, this paper presents an information uniqueness driven video compression framework, termed UniComp, which aims to maximize the information fidelity of video representations under constrained computational budgets. Starting from the information-theoretic perspective, we formulate the vision compression as an optimization problem that minimizes conditional entropy (reconstruction error) between retained and full tokens. To achieve this, we introduce the notion of information uniqueness to measure intrinsic redundancy among tokens to link with reconstruction error. Based on uniqueness, we design three modules-Frame Group Fusion, Token Allocation, and Spatial Dynamic Compression-that progressively perform semantic frame grouping, adaptive resource allocation, and fine-grained spatial compression. Extensive experiments demonstrate that UniComp consistently outperforms existing compression methods in preserving essential visual tokens under limited computational budgets, highlighting the pivotal role of information uniqueness in token compression efficacy.

Paperid: 769, https://arxiv.org/pdf/2512.02399.pdf

Abstract:
Smart contracts are a core component of blockchain technology and are widely deployed across various scenarios. However, atomicity violations have become a potential security risk. Existing analysis tools often lack the precision required to detect these issues effectively. To address this challenge, we introduce AtomGraph, an automated framework designed for detecting atomicity violations. This framework leverages Graph Convolutional Networks (GCN) to identify atomicity violations through multimodal feature learning and fusion. Specifically, driven by a collaborative learning mechanism, the model simultaneously learns from two heterogeneous modalities: extracting structural topological features from the contract's Control Flow Graph (CFG) and uncovering deep semantics from its opcode sequence. We designed an adaptive weighted fusion mechanism to dynamically adjust the weights of features from each modality to achieve optimal feature fusion. Finally, GCN detects graph-level atomicity violation on the contract. Comprehensive experimental evaluations demonstrate that AtomGraph achieves 96.88% accuracy and 96.97% F1 score, outperforming existing tools. Furthermore, compared to the concatenation fusion model, AtomGraph improves the F1 score by 6.4%, proving its potential in smart contract security detection.

Paperid: 770, https://arxiv.org/pdf/2512.01979.pdf

Abstract:
GUI grounding aims to align natural language instructions with precise regions in complex user interfaces. Advanced multimodal large language models show strong ability in visual GUI grounding but still struggle with small or visually similar targets and ambiguity in real world layouts. These limitations arise from limited grounding capacity and from underuse of existing reasoning potential. We present Chain of Ground CoG a training free multi step grounding framework that uses multimodal large language models for iterative visual reasoning and refinement. Instead of direct prediction the model progressively reflects and adjusts its hypotheses leading to more accurate and interpretable localization. Our approach achieves 68.4 accuracy on the ScreenSpot Pro benchmark an improvement of 4.8 points. To measure real world generalization we introduce TPanel UI a dataset of 420 labeled industrial control panels with visual distortions such as blur and masking. On TPanel UI Chain of Ground improves over the strong baseline Qwen3 VL 235B by 6.9 points showing the effectiveness of multi step training free grounding across real world and digital interfaces. These results highlight a direction for unlocking grounding potential through structured iterative refinement instead of additional training.

Paperid: 771, https://arxiv.org/pdf/2512.01577.pdf

Abstract:
Non-fungible tokens (NFTs) serve as a representative form of digital asset ownership and have attracted numerous investors, creators, and tech enthusiasts in recent years. However, related fraud activities, especially phishing scams, have caused significant property losses. There are many graph analysis methods to detect malicious scam incidents, but no research on the transaction patterns of the NFT scams. Therefore, to fill this gap, we are the first to systematically explore NFT phishing frauds through graph analysis, aiming to comprehensively investigate the characteristics and patterns of NFT phishing frauds on the transaction graph. During the research process, we collect transaction records, log data, and security reports related to NFT phishing incidents published on multiple platforms. After collecting, sanitizing, and unifying the data, we construct a transaction graph and analyze the distribution, transaction features, and interaction patterns of NFT phishing scams. We find that normal transactions on the blockchain accounted for 96.71% of all transactions. Although phishing-related accounts accounted for only 0.94% of the total accounts, they appeared in 8.36% of the transaction scenarios, and their interaction probability with normal accounts is significantly higher in large-scale transaction networks. Moreover, NFT phishing scammers often carry out fraud in a collective manner, targeting specific accounts, tend to interact with victims through multiple token standards, have shorter transaction cycles than normal transactions, and involve more multi-party transactions. This study reveals the core behavioral features of NFT phishing scams, providing important references for the detection and prevention of NFT phishing scams in the future.

Paperid: 772, https://arxiv.org/pdf/2512.01030.pdf

Abstract:
Recovering pixel-wise geometric properties from a single image is fundamentally ill-posed due to appearance ambiguity and non-injective mappings between 2D observations and 3D structures. While discriminative regression models achieve strong performance through large-scale supervision, their success is bounded by the scale, quality and diversity of available data and limited physical reasoning. Recent diffusion models exhibit powerful world priors that encode geometry and semantics learned from massive image-text data, yet directly reusing their stochastic generative formulation is suboptimal for deterministic geometric inference: the former is optimized for diverse and high-fidelity image generation, whereas the latter requires stable and accurate predictions. In this work, we propose Lotus-2, a two-stage deterministic framework for stable, accurate and fine-grained geometric dense prediction, aiming to provide an optimal adaption protocol to fully exploit the pre-trained generative priors. Specifically, in the first stage, the core predictor employs a single-step deterministic formulation with a clean-data objective and a lightweight local continuity module (LCM) to generate globally coherent structures without grid artifacts. In the second stage, the detail sharpener performs a constrained multi-step rectified-flow refinement within the manifold defined by the core predictor, enhancing fine-grained geometry through noise-free deterministic flow matching. Using only 59K training samples, less than 1% of existing large-scale datasets, Lotus-2 establishes new state-of-the-art results in monocular depth estimation and highly competitive surface normal prediction. These results demonstrate that diffusion models can serve as deterministic world priors, enabling high-quality geometric reasoning beyond traditional discriminative and generative paradigms.

Paperid: 773, https://arxiv.org/pdf/2512.00966.pdf

Abstract:
Indirect prompt injection attacks (IPIAs), where large language models (LLMs) follow malicious instructions hidden in input data, pose a critical threat to LLM-powered agents. In this paper, we present IntentGuard, a general defense framework based on instruction-following intent analysis. The key insight of IntentGuard is that the decisive factor in IPIAs is not the presence of malicious text, but whether the LLM intends to follow instructions from untrusted data. Building on this insight, IntentGuard leverages an instruction-following intent analyzer (IIA) to identify which parts of the input prompt the model recognizes as actionable instructions, and then flag or neutralize any overlaps with untrusted data segments. To instantiate the framework, we develop an IIA that uses three "thinking intervention" strategies to elicit a structured list of intended instructions from reasoning-enabled LLMs. These techniques include start-of-thinking prefilling, end-of-thinking refinement, and adversarial in-context demonstration. We evaluate IntentGuard on two agentic benchmarks (AgentDojo and Mind2Web) using two reasoning-enabled LLMs (Qwen-3-32B and gpt-oss-20B). Results demonstrate that IntentGuard achieves (1) no utility degradation in all but one setting and (2) strong robustness against adaptive prompt injection attacks (e.g., reducing attack success rates from 100% to 8.5% in a Mind2Web scenario).

Paperid: 774, https://arxiv.org/pdf/2512.00956.pdf

Abstract:
Quantization to low bitwidth is a standard approach for deploying large language models, however, a few extreme weights and activations stretch the dynamic range and reduce the effective resolution of the quantizer. A common mitigation approach is to apply some fixed orthogonal transforms, such as Hadamard matrices, before quantization, which typically reduces the dynamic range. Yet, these transforms ignore the statistics of the data, and their optimality is currently not understood. In this work, we derive, for the first time, closed-form optimal linear blockwise transforms for joint weight-activation quantization using standard data-free quantizers for common numerical formats. Specifically, we provide derivations of the optimal adaptive (data-aware) transforms for round-to-nearest (RTN), AbsMax-scaled block quantizers for both integer and floating-point formats. The resulting construction, which we call WUSH, combines a Hadamard backbone with a data-dependent component based on second-order moments, yielding a non-orthogonal transform that is provably optimal under mild assumptions and remains structured for efficient implementation. Preliminary experimental results show that our approach consistently improves upon the Hadamard transform for common formats.

Paperid: 775, https://arxiv.org/pdf/2512.00944.pdf

Abstract:
3D Gaussian Splatting (3D-GS) has emerged as an efficient 3D representation and a promising foundation for semantic tasks like segmentation. However, existing 3D-GS-based segmentation methods typically rely on high-dimensional category features, which introduce substantial memory overhead. Moreover, fine-grained segmentation remains challenging due to label space congestion and the lack of stable multi-granularity control mechanisms. To address these limitations, we propose a coarse-to-fine binary encoding scheme for per-Gaussian category representation, which compresses each feature into a single integer via the binary-to-decimal mapping, drastically reducing memory usage. We further design a progressive training strategy that decomposes panoptic segmentation into a series of independent sub-tasks, reducing inter-class conflicts and thereby enhancing fine-grained segmentation capability. Additionally, we fine-tune opacity during segmentation training to address the incompatibility between photometric rendering and semantic segmentation, which often leads to foreground-background confusion. Extensive experiments on multiple benchmarks demonstrate that our method achieves state-of-the-art segmentation performance while significantly reducing memory consumption and accelerating inference.

Paperid: 776, https://arxiv.org/pdf/2512.00417.pdf

Abstract:
This paper introduces CryptoBench, the first expert-curated, dynamic benchmark designed to rigorously evaluate the real-world capabilities of Large Language Model (LLM) agents in the uniquely demanding and fast-paced cryptocurrency domain. Unlike general-purpose agent benchmarks for search and prediction, professional crypto analysis presents specific challenges: \emph{extreme time-sensitivity}, \emph{a highly adversarial information environment}, and the critical need to synthesize data from \emph{diverse, specialized sources}, such as on-chain intelligence platforms and real-time Decentralized Finance (DeFi) dashboards. CryptoBench thus serves as a much more challenging and valuable scenario for LLM agent assessment. To address these challenges, we constructed a live, dynamic benchmark featuring 50 questions per month, expertly designed by crypto-native professionals to mirror actual analyst workflows. These tasks are rigorously categorized within a four-quadrant system: Simple Retrieval, Complex Retrieval, Simple Prediction, and Complex Prediction. This granular categorization enables a precise assessment of an LLM agent's foundational data-gathering capabilities alongside its advanced analytical and forecasting skills. Our evaluation of ten LLMs, both directly and within an agentic framework, reveals a performance hierarchy and uncovers a failure mode. We observe a \textit{retrieval-prediction imbalance}, where many leading models, despite being proficient at data retrieval, demonstrate a pronounced weakness in tasks requiring predictive analysis. This highlights a problematic tendency for agents to appear factually grounded while lacking the deeper analytical capabilities to synthesize information.

Paperid: 777, https://arxiv.org/pdf/2512.00345.pdf

Abstract:
Existing Human Motion Prediction (HMP) methods based on RGB-D cameras are sensitive to lighting conditions and raise privacy concerns, limiting their real-world applications such as firefighting and healthcare. Motivated by the robustness and privacy-preserving nature of millimeter-wave (mmWave) radar, this work introduces radar as a novel sensing modality for HMP, for the first time. Nevertheless, radar signals often suffer from specular reflections and multipath effects, resulting in noisy and temporally inconsistent measurements, such as body-part miss-detection. To address these radar-specific artifacts, we propose mmPred, the first diffusion-based framework tailored for radar-based HMP. mmPred introduces a dual-domain historical motion representation to guide the generation process, combining a Time-domain Pose Refinement (TPR) branch for learning fine-grained details and a Frequency-domain Dominant Motion (FDM) branch for capturing global motion trends and suppressing frame-level inconsistency. Furthermore, we design a Global Skeleton-relational Transformer (GST) as the diffusion backbone to model global inter-joint cooperation, enabling corrupted joints to dynamically aggregate information from others. Extensive experiments show that mmPred achieves state-of-the-art performance, outperforming existing methods by 8.6% on mmBody and 22% on mm-Fi.

Paperid: 778, https://arxiv.org/pdf/2512.07165.pdf

Abstract:
Sparse-view 3D Gaussian splatting seeks to render high-quality novel views of 3D scenes from a limited set of input images. While recent pose-free feed-forward methods leveraging pre-trained 3D priors have achieved impressive results, most of them rely on full fine-tuning of large Vision Transformer (ViT) backbones and incur substantial GPU costs. In this work, we introduce MuSASplat, a novel framework that dramatically reduces the computational burden of training pose-free feed-forward 3D Gaussian splats models with little compromise of rendering quality. Central to our approach is a lightweight Multi-Scale Adapter that enables efficient fine-tuning of ViT-based architectures with only a small fraction of training parameters. This design avoids the prohibitive GPU overhead associated with previous full-model adaptation techniques while maintaining high fidelity in novel view synthesis, even with very sparse input views. In addition, we introduce a Feature Fusion Aggregator that integrates features across input views effectively and efficiently. Unlike widely adopted memory banks, the Feature Fusion Aggregator ensures consistent geometric integration across input views and meanwhile mitigates the memory usage, training complexity, and computational costs significantly. Extensive experiments across diverse datasets show that MuSASplat achieves state-of-the-art rendering quality but has significantly reduced parameters and training resource requirements as compared with existing methods.

Paperid: 779, https://arxiv.org/pdf/2512.07034.pdf

Abstract:
Glass is a prevalent material among solid objects in everyday life, yet segmentation methods struggle to distinguish it from opaque materials due to its transparency and reflection. While it is known that human perception relies on boundary and reflective-object features to distinguish glass objects, the existing literature has not yet sufficiently captured both properties when handling transparent objects. Hence, we propose incorporating both of these powerful visual cues via the Boundary Feature Enhancement and Reflection Feature Enhancement modules in a mutually beneficial way. Our proposed framework, TransCues, is a pyramidal transformer encoder-decoder architecture to segment transparent objects. We empirically show that these two modules can be used together effectively, improving overall performance across various benchmark datasets, including glass object semantic segmentation, mirror object semantic segmentation, and generic segmentation datasets. Our method outperforms the state-of-the-art by a large margin, achieving +4.2% mIoU on Trans10K-v2, +5.6% mIoU on MSD, +10.1% mIoU on RGBD-Mirror, +13.1% mIoU on TROSD, and +8.3% mIoU on Stanford2D3D, showing the effectiveness of our method against glass objects.

Paperid: 780, https://arxiv.org/pdf/2512.06494.pdf

Abstract:
This letter describes how to improve performance of OFDM systems by combining non-equiprobable signaling with low density parity check (LDPC) coding. We partition a standard QAM constellation into annular subconstellations of equal size, and we implement non-equiprobable signaling through a shaping code which selects subconstellations with large average energy less frequently than subconstellations with small average energy. In equiprobable signaling, the LDPC code selects a signal point from the inner subconstellation. In non-equiprobable signaling this inner signal point has a representative in each subconstellation and the shaping code selects the representative for transmission. It is possible to use standard QAM constellations to achieve any desired fractional bit rate with this method of shaping the energy distribution of the transmitted signal. We describe how to combine coding and shaping by integrating shaping into the calculation of log-likelihood ratios (LLRs) necessary for decoding LDPC codes. We present simulation results for non-equiprobable transmission at $1.5$ bits/symbol on a representative Veh-A channel showing gains of $4$ dB at a bit error rate (BER) of $10^{-3}$. As the transmission rate increases, the gains from non-equiprobable signaling diminish, but we show through simulation that they are still significant for $16$-QAM.

Paperid: 781, https://arxiv.org/pdf/2512.06287.pdf

Abstract:
In this paper, we develop a hybrid prediction framework for accurate electric vehicle (EV) charging time estimation, a capability that is critical for trip planning, user satisfaction, and efficient operation of charging infrastructure. We combine a physics-based analytical model with a reinforcement learning (RL) approach. The analytical component captures the nonlinear constant-current/constant-voltage (CC--CV) charging dynamics and explicitly models state-of-health (SoH)--dependent capacity and power fade, providing a reliable baseline when historical data are limited. Building on this foundation, we introduce an RL component that progressively refines charging-time predictions as operational data accumulate, enabling improved long-term adaptation. Both models incorporate SoH degradation to maintain predictive accuracy over the battery lifetime. We evaluate the framework using $5{,}000$ simulated charging sessions calibrated to manufacturer specifications and publicly available EV charging datasets. Our results show that the analytical model achieves $R^{2}=98.5\%$ and $\mathrm{MAPE}=2.1\%$, while the RL model further improves performance to $R^{2}=99.2\%$ and $\mathrm{MAPE}=1.6\%$, corresponding to a $23\%$ accuracy gain and $35\%$ improved robustness to battery aging.

Paperid: 782, https://arxiv.org/pdf/2512.06283.pdf

Abstract:
In this paper, we investigate cooperative platoon formation and benefit allocation in mixed-energy truck fleets composed of both electric and fuel-powered trucks. The central challenge arises from the platoon-size constraint, which limits the number of trucks permitted in each platoon and introduces combinatorial coupling into the search for optimal platoon formation structures. We formulate this problem as a coalitional game with bounded coalition sizes and derive a closed-form characterization of the optimal coalition structure that maximizes the fleet-wide platooning benefit. Building on this structure, we develop a type-based least-core payoff allocation scheme that guarantees stability within the coalition-structure core (CS-core). For cases in which the CS-core is empty, we compute the least-core radius to determine the minimal relaxation required to achieve approximate stability. Through numerical studies, we demonstrate that the proposed framework consistently achieves the highest total platooning benefit among all feasible formation configurations while providing stable benefit allocations that outperform existing baseline methods.

Paperid: 783, https://arxiv.org/pdf/2512.05788.pdf

Abstract:
The success of collaborative task completion among networked devices hinges on the effective selection of trustworthy collaborators. However, accurate task-specific trust evaluation of multi-hop collaborators can be extremely complex. The reason is that their trust evaluation is determined by a combination of diverse trust-related perspectives with different characteristics, including historical collaboration reliability, volatile and sensitive conditions of available resources for collaboration, as well as continuously evolving network topologies. To address this challenge, this paper presents a graph neural network (GNN)-aided distributed agentic AI (GADAI) framework, in which different aspects of devices' task-specific trustworthiness are separately evaluated and jointly integrated to facilitate multi-hop collaborator selection. GADAI first utilizes a GNN-assisted model to infer device trust from historical collaboration data. Specifically, it employs GNN to propagate and aggregate trust information among multi-hop neighbours, resulting in more accurate device reliability evaluation. Considering the dynamic and privacy-sensitive nature of device resources, a privacy-preserving resource evaluation mechanism is implemented using agentic AI. Each device hosts a large AI model-driven agent capable of autonomously determining whether its local resources meet the requirements of a given task, ensuring both task-specific and privacy-preserving trust evaluation. By combining the outcomes of these assessments, only the trusted devices can coordinate a task-oriented multi-hop cooperation path through their agents in a distributed manner. Experimental results show that our proposed GADAI outperforms the comparison algorithms in planning multi-hop paths that maximize the value of task completion.

Paperid: 784, https://arxiv.org/pdf/2512.05387.pdf

Abstract:
Large Language Models (LLMs) often suffer from hallucinations: output content that is not grounded in the input context, when performing long-form text generation tasks such as summarization. Prior works have shown that hallucinations can be reduced by iteratively critiquing and refining previously generated outputs using either the same model or a more powerful teacher model as the critique. However, these approaches either require additional test-time compute or assume access to more powerful teacher models, making them costly and less practical. In this work, we propose Self Critique and Refinement-based Preference Optimization (SCRPO), which is a self-supervised training framework that first constructs a preference dataset by leveraging the LLM's own critique and refinement capabilities, and then applies preference learning to improve the same LLM for faithful summarization. Experiments on three summarization benchmarks (XSUM CNNDM and SAMSum), demonstrate that our approach outperforms state-of-the-art self-supervised learning methods in terms of faithfulness metrics while either maintaining or improving other metrics that measure the overall quality of the summary. Moreover, compared to test-time refinement, our approach not only improves efficiency but also results in more faithful summaries.

Paperid: 785, https://arxiv.org/pdf/2512.05152.pdf

Abstract:
Diffusion models are highly regarded for their controllability and the diversity of images they generate. However, class-conditional generation methods based on diffusion models often focus on more common categories. In large-scale fine-grained image generation, issues of semantic information entanglement and insufficient detail in the generated images still persist. This paper attempts to introduce a concept of a tiered embedder in fine-grained image generation, which integrates semantic information from both super and child classes, allowing the diffusion model to better incorporate semantic information and address the issue of semantic entanglement. To address the issue of insufficient detail in fine-grained images, we introduce the concept of super-resolution during the perceptual information generation stage, enhancing the detailed features of fine-grained images through enhancement and degradation models. Furthermore, we propose an efficient ProAttention mechanism that can be effectively implemented in the diffusion model. We evaluate our method through extensive experiments on public benchmarks, demonstrating that our approach outperforms other state-of-the-art fine-tuning methods in terms of performance.

Paperid: 786, https://arxiv.org/pdf/2512.03963.pdf

Abstract:
Enhancing the temporal understanding of Multimodal Large Language Models (MLLMs) is essential for advancing long-form video analysis, enabling tasks such as temporal localization, action detection, and time-sensitive question answering. While reinforcement learning (RL) has recently been explored for improving temporal reasoning, existing approaches are often confined to limited task types and data, restricting their generalization across diverse temporal understanding scenarios. To address this challenge, we present TempR1, a temporal-aware multi-task reinforcement learning framework that systematically strengthens MLLMs' temporal comprehension. We curate a multi-task corpus that exposes the model to diverse temporal structures and semantics, and build upon the Group Relative Policy Optimization (GRPO) algorithm to achieve stable and effective cross-task optimization. Specifically, we categorize temporal tasks into three correspondence types between predicted intervals and ground-truth instances, and design tailored localization rewards for each, enabling TempR1 to capture fine-grained temporal dependencies and adapt to different temporal patterns. Extensive experiments demonstrate that TempR1 attains state-of-the-art performance across multiple benchmarks. Moreover, its joint optimization over complementary tasks yields a strong synergistic effect, enhancing both generalization and single-task performance, establishing a scalable and principled paradigm for temporal reasoning in MLLMs.

Paperid: 787, https://arxiv.org/pdf/2512.03574.pdf

Abstract:
Scene Text Editing (STE) involves replacing text in a scene image with new target text while preserving both the original text style and background texture. Existing methods suffer from two major challenges: inconsistency and length-insensitivity. They often fail to maintain coherence between the edited local patch and the surrounding area, and they struggle to handle significant differences in text length before and after editing. To tackle these challenges, we propose an end-to-end framework called Global-Local Aware Scene Text Editing (GLASTE), which simultaneously incorporates high-level global contextual information along with delicate local features. Specifically, we design a global-local combination structure, joint global and local losses, and enhance text image features to ensure consistency in text style within local patches while maintaining harmony between local and global areas. Additionally, we express the text style as a vector independent of the image size, which can be transferred to target text images of various sizes. We use an affine fusion to fill target text images into the editing patch while maintaining their aspect ratio unchanged. Extensive experiments on real-world datasets validate that our GLASTE model outperforms previous methods in both quantitative metrics and qualitative results and effectively mitigates the two challenges.

Paperid: 788, https://arxiv.org/pdf/2512.03566.pdf

Abstract:
Articulated object generation has seen increasing advancements, yet existing models often lack the ability to be conditioned on text prompts. To address the significant gap between textual descriptions and 3D articulated object representations, we propose GAOT, a three-phase framework that generates articulated objects from text prompts, leveraging diffusion models and hypergraph learning in a three-step process. First, we fine-tune a point cloud generation model to produce a coarse representation of objects from text prompts. Given the inherent connection between articulated objects and graph structures, we design a hypergraph-based learning method to refine these coarse representations, representing object parts as graph vertices. Finally, leveraging a diffusion model, the joints of articulated objects-represented as graph edges-are generated based on the object parts. Extensive qualitative and quantitative experiments on the PartNet-Mobility dataset demonstrate the effectiveness of our approach, achieving superior performance over previous methods.

Paperid: 789, https://arxiv.org/pdf/2512.03540.pdf

Abstract:
Cooking is a sequential and visually grounded activity, where each step such as chopping, mixing, or frying carries both procedural logic and visual semantics. While recent diffusion models have shown strong capabilities in text-to-image generation, they struggle to handle structured multi-step scenarios like recipe illustration. Additionally, current recipe illustration methods are unable to adjust to the natural variability in recipe length, generating a fixed number of images regardless of the actual instructions structure. To address these limitations, we present CookAnything, a flexible and consistent diffusion-based framework that generates coherent, semantically distinct image sequences from textual cooking instructions of arbitrary length. The framework introduces three key components: (1) Step-wise Regional Control (SRC), which aligns textual steps with corresponding image regions within a single denoising process; (2) Flexible RoPE, a step-aware positional encoding mechanism that enhances both temporal coherence and spatial diversity; and (3) Cross-Step Consistency Control (CSCC), which maintains fine-grained ingredient consistency across steps. Experimental results on recipe illustration benchmarks show that CookAnything performs better than existing methods in training-based and training-free settings. The proposed framework supports scalable, high-quality visual synthesis of complex multi-step instructions and holds significant potential for broad applications in instructional media, and procedural content creation.

Paperid: 790, https://arxiv.org/pdf/2512.03247.pdf

Abstract:
Latent Diffusion Models (LDMs) have markedly advanced the quality of image inpainting and local editing. However, the inherent latent compression often introduces pixel-level inconsistencies, such as chromatic shifts, texture mismatches, and visible seams along editing boundaries. Existing remedies, including background-conditioned latent decoding and pixel-space harmonization, usually fail to fully eliminate these artifacts in practice and do not generalize well across different latent representations or tasks. We introduce PixPerfect, a pixel-level refinement framework that delivers seamless, high-fidelity local edits across diverse LDM architectures and tasks. PixPerfect leverages (i) a differentiable discriminative pixel space that amplifies and suppresses subtle color and texture discrepancies, (ii) a comprehensive artifact simulation pipeline that exposes the refiner to realistic local editing artifacts during training, and (iii) a direct pixel-space refinement scheme that ensures broad applicability across diverse latent representations and tasks. Extensive experiments on inpainting, object removal, and insertion benchmarks demonstrate that PixPerfect substantially enhances perceptual fidelity and downstream editing performance, establishing a new standard for robust and high-fidelity localized image editing.

Paperid: 791, https://arxiv.org/pdf/2512.02395.pdf

Abstract:
Despite recent progress in multimodal agentic systems, existing approaches often treat image manipulation and web search as disjoint capabilities, rely heavily on costly reinforcement learning, and lack planning grounded in real tool-execution traces. To address these limitations, we present Skywork-R1V4, a 30B (A3B) parameter multimodal agentic model that unifies multimodal planning, active image manipulation ("thinking with images"), deep multimodal search, and, most critically, interleaved reasoning that dynamically alternates between visual operations and external knowledge retrieval. Trained solely via supervised fine-tuning on fewer than 30,000 high-quality, planning-execution-consistent trajectories and validated through stepwise consistency filtering, Skywork-R1V4 achieves state-of-the-art results across perception and multimodal search benchmarks: it scores 66.1 on MMSearch and 67.2 on FVQA, surpassing Gemini 2.5 Flash on all 11 metrics. Skywork-R1V4 exhibits emergent long-horizon reasoning at inference time, successfully orchestrating more than 10 tool calls to solve complex, multi-step tasks. Our results demonstrate that sophisticated agentic multimodal intelligence can be achieved through carefully curated supervised learning alone, without any reliance on reinforcement learning.

Paperid: 792, https://arxiv.org/pdf/2512.01952.pdf

Abstract:
Recent advances in video world modeling have enabled large-scale generative models to simulate embodied environments with high visual fidelity, providing strong priors for prediction, planning, and control. Yet, despite their realism, these models often lack geometric grounding, limiting their use in navigation tasks that require spatial coherence and long-horizon stability. We introduce Reinforcement Learning with World Grounding (RLWG), a self-supervised post-training framework that aligns pretrained world models with a physically verifiable structure through geometric and perceptual rewards. Analogous to reinforcement learning from verifiable feedback (RLVR) in language models, RLWG can use multiple rewards that measure pose cycle-consistency, depth reprojection, and temporal coherence. We instantiate this framework with GrndCtrl, a reward-aligned adaptation method based on Group Relative Policy Optimization (GRPO), yielding world models that maintain stable trajectories, consistent geometry, and reliable rollouts for embodied navigation. Like post-training alignment in large language models, GrndCtrl leverages verifiable rewards to bridge generative pretraining and grounded behavior, achieving superior spatial coherence and navigation stability over supervised fine-tuning in outdoor environments.

Paperid: 793, https://arxiv.org/pdf/2512.01061.pdf

Abstract:
Recent progress in GPU-accelerated, photorealistic simulation has opened a scalable data-generation path for robot learning, where massive physics and visual randomization allow policies to generalize beyond curated environments. Building on these advances, we develop a teacher-student-bootstrap learning framework for vision-based humanoid loco-manipulation, using articulated-object interaction as a representative high-difficulty benchmark. Our approach introduces a staged-reset exploration strategy that stabilizes long-horizon privileged-policy training, and a GRPO-based fine-tuning procedure that mitigates partial observability and improves closed-loop consistency in sim-to-real RL. Trained entirely on simulation data, the resulting policy achieves robust zero-shot performance across diverse door types and outperforms human teleoperators by up to 31.7% in task completion time under the same whole-body control stack. This represents the first humanoid sim-to-real policy capable of diverse articulated loco-manipulation using pure RGB perception.

Paperid: 794, https://arxiv.org/pdf/2512.07728.pdf

Abstract:
The continuous Frechet distance between two polygonal curves is classically computed by exploring their free space diagram. Recently, Har-Peled, Raichel, and Robson [SoCG'25] proposed a radically different approach: instead of directly traversing the continuous free space, they approximate the distance by computing paths in a discrete graph derived from the discrete free space, recursively bisecting edges until the discrete value converges to the continuous Frechet distance. They implement this so-called frog-based technique and report substantial practical speedups over the state of the art. We revisit the frog-based approach and address three of its limitations. First, the method does not compute the Frechet distance exactly. Second, the recursive bisection procedure only introduces the monotonicity events required to realise the Frechet distance asymptotically, that is, only in the limit. Third, the applied simplification technique is heuristic. Motivated by theoretical considerations, we develop new techniques that guarantee exactness, polynomial-time convergence, and near-optimal lossless simplifications. We provide an open-source C++ implementation of our variant. Our primary contribution is an extensive empirical evaluation. As expected, exact computation introduces overhead and increases the median running time. Yet, method is often faster in the worst case, the slowest ten percent of instances, or even on average due to its convergence guarantees. More surprisingly, in our experiments, the implementation of Bringmann, Kuennemann, and Nusser [SoCG'19] consistently outperforms all frog-based approaches in practice. This appears to contrast published claims of the efficiency of the frog-based techniques. These results thereby provide nuanced perspective on frogs: highlighting both the theoretical appeal, but also the practical limitations.

Paperid: 795, https://arxiv.org/pdf/2512.07251.pdf

Abstract:
Image enhancement improves visual quality and helps reveal details that are hard to see in the original image. In medical imaging, it can support clinical decision-making, but current models often over-edit. This can distort organs, create false findings, and miss small tumors because these models do not understand anatomy or contrast dynamics. We propose SMILE, an anatomy-aware diffusion model that learns how organs are shaped and how they take up contrast. It enhances only clinically relevant regions while leaving all other areas unchanged. SMILE introduces three key ideas: (1) structure-aware supervision that follows true organ boundaries and contrast patterns; (2) registration-free learning that works directly with unaligned multi-phase CT scans; (3) unified inference that provides fast and consistent enhancement across all contrast phases. Across six external datasets, SMILE outperforms existing methods in image quality (14.2% higher SSIM, 20.6% higher PSNR, 50% better FID) and in clinical usefulness by producing anatomically accurate and diagnostically meaningful images. SMILE also improves cancer detection from non-contrast CT, raising the F1 score by up to 10 percent.

Paperid: 796, https://arxiv.org/pdf/2512.06559.pdf

Abstract:
TimSort is a well-established sorting algorithm whose running time depends on how sorted the input already is. Recently, Eppstein, Goodrich, Illickan, and To designed algorithms inspired by TimSort for Pareto front, planar convex hull, and two other problems. For each of these problems, they define a Range Partition Entropy; a function $H$ mapping lists $I$ that store $n$ points to a number between $0$ and $\log n$. Their algorithms have, for each list of points $I$, a running time of $O(n(1 + H(I)))$. In this paper, we provide matching lower bounds for the Pareto front and convex hull algorithms by Eppstein, Goodrich, Illickan, and To. In particular, we show that their algorithm does not correspond to TimSort (or related stack-based MergeSort variants) but rather to a variant of QuickSort. From this, we derive an intuitive notion of universal optimality. We show comparison-based lower bounds that prove that the algorithms by Eppstein, Goodrich, Illickan and To are universally optimal under this notion of universal optimality.

Paperid: 797, https://arxiv.org/pdf/2512.06251.pdf

Abstract:
Partially Supervised Multi-Task Learning (PS-MTL) aims to leverage knowledge across tasks when annotations are incomplete. Existing approaches, however, have largely focused on the simpler setting of homogeneous, dense prediction tasks, leaving the more realistic challenge of learning from structurally diverse tasks unexplored. To this end, we introduce NexusFlow, a novel, lightweight, and plug-and-play framework effective in both settings. NexusFlow introduces a set of surrogate networks with invertible coupling layers to align the latent feature distributions of tasks, creating a unified representation that enables effective knowledge transfer. The coupling layers are bijective, preserving information while mapping features into a shared canonical space. This invertibility avoids representational collapse and enables alignment across structurally different tasks without reducing expressive capacity. We first evaluate NexusFlow on the core challenge of domain-partitioned autonomous driving, where dense map reconstruction and sparse multi-object tracking are supervised in different geographic regions, creating both structural disparity and a strong domain gap. NexusFlow sets a new state-of-the-art result on nuScenes, outperforming strong partially supervised baselines. To demonstrate generality, we further test NexusFlow on NYUv2 using three homogeneous dense prediction tasks, segmentation, depth, and surface normals, as a representative N-task PS-MTL scenario. NexusFlow yields consistent gains across all tasks, confirming its broad applicability.

Paperid: 798, https://arxiv.org/pdf/2512.05428.pdf

Abstract:
Bias in AI systems can lead to unfair and discriminatory outcomes, especially when left untested before deployment. Although fairness testing aims to identify and mitigate such bias, existing tools are often difficult to use, requiring advanced expertise and offering limited support for real-world workflows. To address this, we introduce Bita, a conversational assistant designed to help software testers detect potential sources of bias, evaluate test plans through a fairness lens, and generate fairness-oriented exploratory testing charters. Bita integrates a large language model with retrieval-augmented generation, grounding its responses in curated fairness literature. Our validation demonstrates how Bita supports fairness testing tasks on real-world AI systems, providing structured, reproducible evidence of its utility. In summary, our work contributes a practical tool that operationalizes fairness testing in a way that is accessible, systematic, and directly applicable to industrial practice.

Paperid: 799, https://arxiv.org/pdf/2512.05412.pdf

Abstract:
Manual pruning of radiata pine trees poses significant safety risks due to extreme working heights and challenging terrain. This paper presents a computer vision framework that integrates YOLO object detection with Semi-Global Block Matching (SGBM) stereo vision for autonomous drone-based pruning operations. Our system achieves precise branch detection and depth estimation using only stereo camera input, eliminating the need for expensive LiDAR sensors. Experimental evaluation demonstrates YOLO's superior performance over Mask R-CNN, achieving 82.0% mAPmask50-95 for branch segmentation. The integrated system accurately localizes branches within a 2 m operational range, with processing times under one second per frame. These results establish the feasibility of cost-effective autonomous pruning systems that enhance worker safety and operational efficiency in commercial forestry.

Paperid: 800, https://arxiv.org/pdf/2512.05410.pdf

Abstract:
Traditional stereo matching algorithms like Semi-Global Block Matching (SGBM) with Weighted Least Squares (WLS) filtering offer speed advantages over neural networks for UAV applications, generating disparity maps in approximately 0.5 seconds per frame. However, these algorithms require meticulous parameter tuning. We propose a Genetic Algorithm (GA) based parameter optimization framework that systematically searches for optimal parameter configurations for SGBM and WLS, enabling UAVs to measure distances to tree branches with enhanced precision while maintaining processing efficiency. Our contributions include: (1) a novel GA-based parameter optimization framework that eliminates manual tuning; (2) a comprehensive evaluation methodology using multiple image quality metrics; and (3) a practical solution for resource-constrained UAV systems. Experimental results demonstrate that our GA-optimized approach reduces Mean Squared Error by 42.86% while increasing Peak Signal-to-Noise Ratio and Structural Similarity by 8.47% and 28.52%, respectively, compared with baseline configurations. Furthermore, our approach demonstrates superior generalization performance across varied imaging conditions, which is critcal for real-world forestry applications.

Paperid: 801, https://arxiv.org/pdf/2512.05136.pdf

Abstract:
Coronary artery disease (CAD) remains a major global health burden. Accurate identification of the culprit vessel and assessment of stenosis severity are essential for guiding individualized therapy. Although coronary CT angiography (CCTA) is the first-line non-invasive modality for CAD diagnosis, its dependence on high-end equipment, radiation exposure, and strict patient cooperation limits large-scale use. With advances in artificial intelligence (AI) and the widespread availability of electrocardiography (ECG), AI-ECG offers a promising alternative for CAD screening. In this study, we developed an interpretable AI-ECG model to predict severe or complete stenosis of the four major coronary arteries on CCTA. On the internal validation set, the model's AUCs for the right coronary artery (RCA), left main coronary artery (LM), left anterior descending artery (LAD), and left circumflex artery (LCX) were 0.794, 0.818, 0.744, and 0.755, respectively; on the external validation set, the AUCs reached 0.749, 0.971, 0.667, and 0.727, respectively. Performance remained stable in a clinically normal-ECG subset, indicating robustness beyond overt ECG abnormalities. Subgroup analyses across demographic and acquisition-time strata further confirmed model stability. Risk stratification based on vessel-specific incidence thresholds showed consistent separation on calibration and cumulative event curves. Interpretability analyses revealed distinct waveform differences between high- and low-risk groups, highlighting key electrophysiological regions contributing to model decisions and offering new insights into the ECG correlates of coronary stenosis.

Paperid: 802, https://arxiv.org/pdf/2512.04829.pdf

Abstract:
Sphere packing, Hilbert's eighteenth problem, asks for the densest arrangement of congruent spheres in n-dimensional Euclidean space. Although relevant to areas such as cryptography, crystallography, and medical imaging, the problem remains unresolved: beyond a few special dimensions, neither optimal packings nor tight upper bounds are known. Even a major breakthrough in dimension $n=8$, later recognised with a Fields Medal, underscores its difficulty. A leading technique for upper bounds, the three-point method, reduces the problem to solving large, high-precision semidefinite programs (SDPs). Because each candidate SDP may take days to evaluate, standard data-intensive AI approaches are infeasible. We address this challenge by formulating SDP construction as a sequential decision process, the SDP game, in which a policy assembles SDP formulations from a set of admissible components. Using a sample-efficient model-based framework that combines Bayesian optimisation with Monte Carlo Tree Search, we obtain new state-of-the-art upper bounds in dimensions $4-16$, showing that model-based search can advance computational progress in longstanding geometric problems. Together, these results demonstrate that sample-efficient, model-based search can make tangible progress on mathematically rigid, evaluation limited problems, pointing towards a complementary direction for AI-assisted discovery beyond large-scale LLM-driven exploration.

Paperid: 803, https://arxiv.org/pdf/2512.04315.pdf

Abstract:
Modeling dynamic 3D scenes is challenging due to their high-dimensional nature, which requires aggregating information from multiple views to reconstruct time-evolving 3D geometry and motion. We present a novel multi-video 4D Gaussian Splatting (4DGS) approach designed to handle real-world, unsynchronized video sets. Our approach, SyncTrack4D, directly leverages dense 4D track representation of dynamic scene parts as cues for simultaneous cross-video synchronization and 4DGS reconstruction. We first compute dense per-video 4D feature tracks and cross-video track correspondences by Fused Gromov-Wasserstein optimal transport approach. Next, we perform global frame-level temporal alignment to maximize overlapping motion of matched 4D tracks. Finally, we achieve sub-frame synchronization through our multi-video 4D Gaussian splatting built upon a motion-spline scaffold representation. The final output is a synchronized 4DGS representation with dense, explicit 3D trajectories, and temporal offsets for each video. We evaluate our approach on the Panoptic Studio and SyncNeRF Blender, demonstrating sub-frame synchronization accuracy with an average temporal error below 0.26 frames, and high-fidelity 4D reconstruction reaching 26.3 PSNR scores on the Panoptic Studio dataset. To the best of our knowledge, our work is the first general 4D Gaussian Splatting approach for unsynchronized video sets, without assuming the existence of predefined scene objects or prior models.

Paperid: 804, https://arxiv.org/pdf/2512.03727.pdf

Abstract:
Probabilistic Graphical Models (PGMs) encode conditional dependencies among random variables using a graph -nodes for variables, links for dependencies- and factorize the joint distribution into lower-dimensional components. This makes PGMs well-suited for analyzing complex systems and supporting decision-making. Recent advances in topological signal processing highlight the importance of variables defined on topological spaces in several application domains. In such cases, the underlying topology shapes statistical relationships, limiting the expressiveness of canonical PGMs. To overcome this limitation, we introduce Colored Markov Random Fields (CMRFs), which model both conditional and marginal dependencies among Gaussian edge variables on topological spaces, with a theoretical foundation in Hodge theory. CMRFs extend classical Gaussian Markov Random Fields by including link coloring: connectivity encodes conditional independence, while color encodes marginal independence. We quantify the benefits of CMRFs through a distributed estimation case study over a physical network, comparing it with baselines with different levels of topological prior.

Paperid: 805, https://arxiv.org/pdf/2512.03427.pdf

Abstract:
Autonomous UAV forestry operations require robust depth estimation methods with strong cross-domain generalization. However, existing evaluations focus on urban and indoor scenarios, leaving a critical gap for specialized vegetation-dense environments. We present the first systematic zero-shot evaluation of eight state-of-the-art stereo methods--RAFT-Stereo, IGEV, IGEV++, BridgeDepth, StereoAnywhere, DEFOM (plus baseline methods ACVNet, PSMNet, TCstereo)--spanning iterative refinement, foundation model, and zero-shot adaptation paradigms. All methods are trained exclusively on Scene Flow and evaluated without fine-tuning on four standard benchmarks (ETH3D, KITTI 2012/2015, Middlebury) plus a novel 5,313-pair Canterbury forestry dataset captured with ZED Mini camera (1920x1080). Performance reveals scene-dependent patterns: foundation models excel on structured scenes (BridgeDepth: 0.23 px on ETH3D, 0.83-1.07 px on KITTI; DEFOM: 0.35-4.65 px across benchmarks), while iterative methods maintain cross-domain robustness (IGEV++: 0.36-6.77 px; IGEV: 0.33-21.91 px). Critical finding: RAFT-Stereo exhibits catastrophic ETH3D failure (26.23 px EPE, 98 percent error rate) due to negative disparity predictions, while performing normally on KITTI (0.90-1.11 px). Qualitative evaluation on Canterbury forestry dataset identifies DEFOM as the optimal gold-standard baseline for vegetation depth estimation, exhibiting superior depth smoothness, occlusion handling, and cross-domain consistency compared to IGEV++, despite IGEV++'s finer detail preservation.

Paperid: 806, https://arxiv.org/pdf/2512.03248.pdf

Abstract:
Recent advances in AI call for a paradigm shift from bit-centric communication to goal- and semantics-oriented architectures, paving the way for AI-native 6G networks. In this context, we address a key open challenge: enabling heterogeneous AI agents to exchange compressed latent-space representations while mitigating semantic noise and preserving task-relevant meaning. We cast this challenge as learning both the communication topology and the alignment maps that govern information exchange among agents, yielding a learned network sheaf equipped with orthogonal maps. This learning process is further supported by a semantic denoising end compression module that constructs a shared global semantic space and derives sparse, structured representations of each agent's latent space. This corresponds to a nonconvex dictionary learning problem solved iteratively with closed-form updates. Experiments with mutiple AI agents pre-trained on real image data show that the semantic denoising and compression facilitates AI agents alignment and the extraction of semantic clusters, while preserving high accuracy in downstream task. The resulting communication network provides new insights about semantic heterogeneity across agents, highlighting the interpretability of our methodology.

Paperid: 807, https://arxiv.org/pdf/2512.03005.pdf

Abstract:
The rapid advancement of large language models (LLMs) has opened new possibilities for AI for good applications. As LLMs increasingly mediate online communication, their potential to foster empathy and constructive dialogue becomes an important frontier for responsible AI research. This work explores whether LLMs can serve not only as moderators that detect harmful content, but as mediators capable of understanding and de-escalating online conflicts. Our framework decomposes mediation into two subtasks: judgment, where an LLM evaluates the fairness and emotional dynamics of a conversation, and steering, where it generates empathetic, de-escalatory messages to guide participants toward resolution. To assess mediation quality, we construct a large Reddit-based dataset and propose a multi-stage evaluation pipeline combining principle-based scoring, user simulation, and human comparison. Experiments show that API-based models outperform open-source counterparts in both reasoning and intervention alignment when doing mediation. Our findings highlight both the promise and limitations of current LLMs as emerging agents for online social mediation.

Paperid: 808, https://arxiv.org/pdf/2512.01610.pdf

Abstract:
Multi-Agent System (MAS) developing frameworks serve as the foundational infrastructure for social simulations powered by Large Language Models (LLMs). However, existing frameworks fail to adequately support large-scale simulation development due to inherent limitations in adaptability, configurability, reliability, and code reusability. For example, they cannot simulate a society where the agent population and profiles change over time. To fill this gap, we propose Agent-Kernel, a framework built upon a novel society-centric modular microkernel architecture. It decouples core system functions from simulation logic and separates cognitive processes from physical environments and action execution. Consequently, Agent-Kernel achieves superior adaptability, configurability, reliability, and reusability. We validate the framework's superiority through two distinct applications: a simulation of the Universe 25 (Mouse Utopia) experiment, which demonstrates the handling of rapid population dynamics from birth to death; and a large-scale simulation of the Zhejiang University Campus Life, successfully coordinating 10,000 heterogeneous agents, including students and faculty.

Paperid: 809, https://arxiv.org/pdf/2512.01336.pdf

Abstract:
Humanoid robots have received significant research interests and advancements in recent years. Despite many successes, due to their morphology, dynamics and limitation of control policy, humanoid robots are prone to fall as compared to other embodiments like quadruped or wheeled robots. And its large weight, tall Center of Mass, high Degree-of-Freedom would cause serious hardware damages when falling uncontrolled, to both itself and surrounding objects. Existing researches in this field mostly focus on using control based methods that struggle to cater diverse falling scenarios and may introduce unsuitable human prior. On the other hand, large-scale Deep Reinforcement Learning and Curriculum Learning could be employed to incentivize humanoid agent discovering falling protection policy that fits its own nature and property. In this work, with carefully designed reward functions and domain diversification curriculum, we successfully train humanoid agent to explore falling protection behaviors and discover that by forming a `triangle' structure, the falling damages could be significantly reduced with its rigid-material body. With comprehensive metrics and experiments, we quantify its performance with comparison to other methods, visualize its falling behaviors and successfully transfer it to real world platform.

Paperid: 810, https://arxiv.org/pdf/2512.00369.pdf

Abstract:
The Inversion-Denoising Paradigm, which is based on diffusion models, excels in diverse image editing and restoration tasks. We revisit its mechanism and reveal a critical, overlooked factor in reconstruction degradation: the approximate noise error. This error stems from approximating the noise at step t with the prediction at step t-1, resulting in severe error accumulation throughout the inversion process. We introduce Projection-Orthogonal Least Squares for Robust and Adaptive Inversion (POLARIS), which reformulates inversion from an error-compensation problem into an error-origin problem. Rather than optimizing embeddings or latent codes to offset accumulated drift, POLARIS treats the guidance scale ω as a step-wise variable and derives a mathematically grounded formula to minimize inversion error at each step. Remarkably, POLARIS improves inversion latent quality with just one line of code. With negligible performance overhead, it substantially mitigates noise approximation errors and consistently improves the accuracy of downstream tasks.

Paperid: 811, https://arxiv.org/pdf/2512.07168.pdf

Abstract:
We introduce a two-stage self-supervised framework that combines the Joint-Embedding Predictive Architecture (JEPA) with a Density Adaptive Attention Mechanism (DAAM) for learning robust speech representations. Stage~1 uses JEPA with DAAM to learn semantic audio features via masked prediction in latent space, fully decoupled from waveform reconstruction. Stage~2 leverages these representations for efficient tokenization using Finite Scalar Quantization (FSQ) and a mixed-radix packing scheme, followed by high-fidelity waveform reconstruction with a HiFi-GAN decoder. By integrating Gaussian mixture-based density-adaptive gating into the JEPA encoder, the model performs adaptive temporal feature selection and discovers hierarchical speech structure at a low frame rate of 2.5~Hz. The resulting tokens (47.5 tokens/sec) provide a reversible, highly compressed, and language-model-friendly representation that is competitive with, and often more efficient than, existing neural audio codecs.

Paperid: 812, https://arxiv.org/pdf/2512.06810.pdf

Abstract:
Recent advances in video multimodal large language models (Video MLLMs) have significantly enhanced video understanding and multi-modal interaction capabilities. While most existing systems operate in a turn-based manner where the model can only reply after user turns, proactively deciding when to reply during video playback presents a promising yet challenging direction for real-time applications. In this work, we propose a novel text-to-text approach to proactive interaction, where the model autonomously determines whether to respond or remain silent at each turn based on dialogue history and visual context up to current frame of an streaming video. To overcome difficulties in previous methods such as manually tuning response decision thresholds and annotating precise reply times, we introduce a multi-turn RL based training method that encourages timely and accurate responses without requiring precise response time annotations. We train our model MMDuet2 on a dataset of 52k videos with two types of dialogues via SFT and RL. Experimental results demonstrate that MMDuet2 outperforms existing proactive Video MLLM baselines in response timing and quality, achieving state-of-the-art performance on the ProactiveVideoQA benchmark.

Paperid: 813, https://arxiv.org/pdf/2512.06628.pdf

Abstract:
Embodied imitation learning is constrained by the scarcity of diverse, long-horizon robotic manipulation data. Existing video generation models for this domain are limited to synthesizing short clips of simple actions and often rely on manually defined trajectories. To this end, we introduce MIND-V, a hierarchical framework designed to synthesize physically plausible and logically coherent videos of long-horizon robotic manipulation. Inspired by cognitive science, MIND-V bridges high-level reasoning with pixel-level synthesis through three core components: a Semantic Reasoning Hub (SRH) that leverages a pre-trained vision-language model for task planning; a Behavioral Semantic Bridge (BSB) that translates abstract instructions into domain-invariant representations; and a Motor Video Generator (MVG) for conditional video rendering. MIND-V employs Staged Visual Future Rollouts, a test-time optimization strategy to enhance long-horizon robustness. To align the generated videos with physical laws, we introduce a GRPO reinforcement learning post-training phase guided by a novel Physical Foresight Coherence (PFC) reward. PFC leverages the V-JEPA world model to enforce physical plausibility by aligning the predicted and actual dynamic evolutions in the feature space. MIND-V demonstrates state-of-the-art performance in long-horizon robotic manipulation video generation, establishing a scalable and controllable paradigm for embodied data synthesis.

Paperid: 814, https://arxiv.org/pdf/2512.05391.pdf

Abstract:
Whole Slide Image (WSI) understanding is fundamentally challenging due to its gigapixel scale and the extreme sparsity of diagnostically relevant regions. Unlike human experts who primarily rely on key areas to arrive at a diagnosis, existing slide-level multimodal large language models (MLLMs) for pathology rely on heavy slide-level encoders that process thousands of patch features in a brute-force manner, resulting in excessive computational cost. In this work, we revisit the WSI-language modeling paradigm and show that tile-level features exhibit strong global and local redundancy, whereas only a small subset of tiles are truly task-relevant. Motivated by this observation, we introduce an efficient MLLM framework, called LoC-Path, that replaces the expensive slide-level encoder with redundancy-reducing modules. We first design a Sparse Token Merger (STM) and an MAE-pretrained resampler to remove local redundancy and compress globally redundant tile tokens into a compact slide-level representation set. We then propose a Cross-Attention Routing Adapter (CARA) and a Token Importance Scorer (TIS) to integrate the compressed visual representation with the language model in a computation-efficient manner. Extensive experiments demonstrate that our approach achieves performance comparable to existing state-of-the-art whole-slide MLLMs, while requiring significantly lower computation and memory.

Paperid: 815, https://arxiv.org/pdf/2512.05371.pdf

Abstract:
While Large Language Models (LLMs) demonstrate immense potential for automating integrated circuit (IC) development, their practical deployment is fundamentally limited by restricted context windows. Existing context-extension methods struggle to achieve effective semantic modeling and thorough multi-hop reasoning over extensive, intricate circuit specifications. To address this, we introduce ChipMind, a novel knowledge graph-augmented reasoning framework specifically designed for lengthy IC specifications. ChipMind first transforms circuit specifications into a domain-specific knowledge graph ChipKG through the Circuit Semantic-Aware Knowledge Graph Construction methodology. It then leverages the ChipKG-Augmented Reasoning mechanism, combining information-theoretic adaptive retrieval to dynamically trace logical dependencies with intent-aware semantic filtering to prune irrelevant noise, effectively balancing retrieval completeness and precision. Evaluated on an industrial-scale specification reasoning benchmark, ChipMind significantly outperforms state-of-the-art baselines, achieving an average improvement of 34.59% (up to 72.73%). Our framework bridges a critical gap between academic research and practical industrial deployment of LLM-aided Hardware Design (LAD).

Paperid: 816, https://arxiv.org/pdf/2512.05145.pdf

Abstract:
Effective judges of Vision-Language Models (VLMs) are crucial for model development. Current methods for training VLM judges mainly rely on large-scale human preference annotations. However, such an approach is costly, and the annotations easily become obsolete as models rapidly improve. In this work, we present a framework to self-train a VLM judge model without any human preference annotations, using only self-synthesized data. Our method is iterative and has three stages: (1) generate diverse multimodal instruction-response pairs at varying quality levels, (2) generate reasoning traces and judgments for each pair, removing the ones that do not match our expected quality levels, and (3) training on correct judge answers and their reasoning traces. We evaluate the resulting judge on Multimodal RewardBench and VL-RewardBench across domains: correctness, preference, reasoning, safety, and visual question-answering. Our method improves a Llama-3.2-11B multimodal judge from 0.38 to 0.51 in overall accuracy on VL-RewardBench, often outperforming much larger models including Llama-3.2-90B, GPT-4o, and Claude 3.5 Sonnet, with particularly strong gains in general, hallucination, and reasoning dimensions. The overall strength of these human-annotation-free results suggest the potential for a future self-judge that evolves alongside rapidly improving VLM capabilities.

Paperid: 817, https://arxiv.org/pdf/2512.04884.pdf

Abstract:
We present a dataset for force-grounded, cross-view articulated manipulation that couples what is seen with what is done and what is felt during real human interaction. The dataset contains 3048 sequences across 381 articulated objects in 38 environments. Each object is operated under four embodiments - (i) human hand, (ii) human hand with a wrist-mounted camera, (iii) handheld UMI gripper, and (iv) a custom Hoi! gripper - where the tool embodiment provides synchronized end-effector forces and tactile sensing. Our dataset offers a holistic view of interaction understanding from video, enabling researchers to evaluate how well methods transfer between human and robotic viewpoints, but also investigate underexplored modalities such as force sensing and prediction.

Paperid: 818, https://arxiv.org/pdf/2512.04678.pdf

Abstract:
Efficient streaming video generation is critical for simulating interactive and dynamic worlds. Existing methods distill few-step video diffusion models with sliding window attention, using initial frames as sink tokens to maintain attention performance and reduce error accumulation. However, video frames become overly dependent on these static tokens, resulting in copied initial frames and diminished motion dynamics. To address this, we introduce Reward Forcing, a novel framework with two key designs. First, we propose EMA-Sink, which maintains fixed-size tokens initialized from initial frames and continuously updated by fusing evicted tokens via exponential moving average as they exit the sliding window. Without additional computation cost, EMA-Sink tokens capture both long-term context and recent dynamics, preventing initial frame copying while maintaining long-horizon consistency. Second, to better distill motion dynamics from teacher models, we propose a novel Rewarded Distribution Matching Distillation (Re-DMD). Vanilla distribution matching treats every training sample equally, limiting the model's ability to prioritize dynamic content. Instead, Re-DMD biases the model's output distribution toward high-reward regions by prioritizing samples with greater dynamics rated by a vision-language model. Re-DMD significantly enhances motion quality while preserving data fidelity. We include both quantitative and qualitative experiments to show that Reward Forcing achieves state-of-the-art performance on standard benchmarks while enabling high-quality streaming video generation at 23.1 FPS on a single H100 GPU.

Paperid: 819, https://arxiv.org/pdf/2512.04581.pdf

Abstract:
Unmanned aerial vehicle (UAV) target tracking based on thermal infrared imaging has been one of the most important sensing technologies in anti-UAV applications. However, the infrared UAV targets often exhibit weak features and complex backgrounds, posing significant challenges to accurate tracking. To address these problems, we introduce SiamDFF, a novel dynamic feature fusion Siamese network that integrates feature enhancement and global contextual attention knowledge distillation for infrared UAV target (IRUT) tracking. The SiamDFF incorporates a selective target enhancement network (STEN), a dynamic spatial feature aggregation module (DSFAM), and a dynamic channel feature aggregation module (DCFAM). The STEN employs intensity-aware multi-head cross-attention to adaptively enhance important regions for both template and search branches. The DSFAM enhances multi-scale UAV target features by integrating local details with global features, utilizing spatial attention guidance within the search frame. The DCFAM effectively integrates the mixed template generated from STEN in the template branch and original template, avoiding excessive background interference with the template and thereby enhancing the emphasis on UAV target region features within the search frame. Furthermore, to enhance the feature extraction capabilities of the network for IRUT without adding extra computational burden, we propose a novel tracking-specific target-aware contextual attention knowledge distiller. It transfers the target prior from the teacher network to the student model, significantly improving the student network's focus on informative regions at each hierarchical level of the backbone network. Extensive experiments on real infrared UAV datasets demonstrate that the proposed approach outperforms state-of-the-art target trackers under complex backgrounds while achieving a real-time tracking speed.

Paperid: 820, https://arxiv.org/pdf/2512.04521.pdf

Abstract:
While fulfilling communication tasks, wireless signals can also be used to sense the environment. Among various types of sensing media, WiFi signals offer advantages such as widespread availability, low hardware cost, and strong robustness to environmental conditions like light, temperature, and humidity. By analyzing Wi-Fi signals in the environment, it is possible to capture dynamic changes of the human body and accomplish sensing applications such as gesture recognition. Although many existing gesture sensing solutions perform well in-domain but lack cross-domain capabilities (i.e., recognition performance in untrained environments). To address this, we extract Doppler spectra from the channel state information (CSI) received by all receivers and concatenate each Doppler spectrum along the same time axis to generate fused images with multi-angle information as input features. Furthermore, inspired by the convolutional block attention module (CBAM), we propose a gesture recognition network that integrates a multi-semantic spatial attention mechanism with a self-attention-based channel mechanism. This network constructs attention maps to quantify the spatiotemporal features of gestures in images, enabling the extraction of key domain-independent features. Additionally, ResNet18 is employed as the backbone network to further capture deep-level features. To validate the network performance, we evaluate the proposed network on the public Widar3 dataset, and the results show that it not only maintains high in-domain accuracy of 99.72%, but also achieves high performance in cross-domain recognition of 97.61%, significantly outperforming existing best solutions.

Paperid: 821, https://arxiv.org/pdf/2512.04211.pdf

Abstract:
Quantum networks are expected to be heterogeneous systems, combining distinct qubit platforms, photon wavelengths, and device timescales to achieve scalable, multiuser connectivity. Building and iterating on such systems is costly and slow, motivating hardware-faithful simulations to explore architecture design space and justify implementation decisions. This paper presents a framework for simulating heterogeneous quantum networks based on SeQUeNCe, a discrete-event simulator of quantum networks. We introduce faithful device models for two representative platforms - Ytterbium atoms and superconducting qubits. On top of these models, we implement entanglement generation and entanglement swapping protocols for time-bin encoded photons that account for disparate clock rates and quantum frequency conversion and transducer losses/noise brought by the heterogeneity. Using extensive simulations, we map the rate-fidelity trade space and identify the dominant bottlenecks unique to heterogeneous systems. The models are open source and extensible, enabling reproducible evaluation of future heterogeneous designs and protocols.

Paperid: 822, https://arxiv.org/pdf/2512.03046.pdf

Abstract:
We propose MagicQuill V2, a novel system that introduces a \textbf{layered composition} paradigm to generative image editing, bridging the gap between the semantic power of diffusion models and the granular control of traditional graphics software. While diffusion transformers excel at holistic generation, their use of singular, monolithic prompts fails to disentangle distinct user intentions for content, position, and appearance. To overcome this, our method deconstructs creative intent into a stack of controllable visual cues: a content layer for what to create, a spatial layer for where to place it, a structural layer for how it is shaped, and a color layer for its palette. Our technical contributions include a specialized data generation pipeline for context-aware content integration, a unified control module to process all visual cues, and a fine-tuned spatial branch for precise local editing, including object removal. Extensive experiments validate that this layered approach effectively resolves the user intention gap, granting creators direct, intuitive control over the generative process.

Paperid: 823, https://arxiv.org/pdf/2512.02841.pdf

Abstract:
System prompts provide a lightweight yet powerful mechanism for conditioning large language models (LLMs) at inference time. While prior work has focused on English-only settings, real-world deployments benefit from having a single prompt to operate reliably across languages. This paper presents a comprehensive study of how different system prompts steer models toward accurate and robust cross-lingual behavior. We propose a unified four-dimensional evaluation framework to assess system prompts in multilingual environments. Through large-scale experiments on five languages, three LLMs, and three benchmarks, we uncover that certain prompt components, such as CoT, emotion, and scenario, correlate with robust multilingual behavior. We develop a prompt optimization framework for multilingual settings and show it can automatically discover prompts that improve all metrics by 5-10%. Finally, we analyze over 10 million reasoning units and find that more performant system prompts induce more structured and consistent reasoning patterns, while reducing unnecessary language-switching. Together, we highlight system prompt optimization as a scalable path to accurate and robust multilingual LLM behavior.

Paperid: 824, https://arxiv.org/pdf/2512.02790.pdf

Abstract:
With the rapid advances of powerful multimodal models such as GPT-4o, Nano Banana, and Seedream 4.0 in Image Editing, the performance gap between closed-source and open-source models is widening, primarily due to the scarcity of large-scale, high-quality training data and comprehensive benchmarks capable of diagnosing model weaknesses across diverse editing behaviors. Existing data construction methods face a scale-quality trade-off: human annotations are high-quality but not scalable, while automated pipelines suffer from error propagation and noise. To address this, we introduce a lightweight data pipeline that replaces multi-toolchains with an end-to-end model and a unified post-verification stage. For scalable quality control, we train a 7B dual-task expert model, \textbf{Qwen-Verify}, for efficient failure detection and instruction recaptioning. This pipeline yields \textbf{UnicEdit-10M}, a 10M-scale dataset spanning diverse basic and complex editing tasks. We also propose \textbf{UnicBench}, a general benchmark that extends beyond basic edits to explicitly assess spatial and knowledge-driven reasoning. To enable fine-grained diagnosis, we introduce novel metrics, including \textit{Non-edit Consistency} and \textit{Reasoning Accuracy}. Our analysis of mainstream models on UnicBench reveals their limitations and provides clear directions for future research.

Paperid: 825, https://arxiv.org/pdf/2512.02486.pdf

Abstract:
Single-domain offline reinforcement learning (RL) often suffers from limited data coverage, while cross-domain offline RL handles this issue by leveraging additional data from other domains with dynamics shifts. However, existing studies primarily focus on train-time robustness (handling dynamics shifts from training data), neglecting the test-time robustness against dynamics perturbations when deployed in practical scenarios. In this paper, we investigate dual (both train-time and test-time) robustness against dynamics shifts in cross-domain offline RL. We first empirically show that the policy trained with cross-domain offline RL exhibits fragility under dynamics perturbations during evaluation, particularly when target domain data is limited. To address this, we introduce a novel robust cross-domain Bellman (RCB) operator, which enhances test-time robustness against dynamics perturbations while staying conservative to the out-of-distribution dynamics transitions, thus guaranteeing the train-time robustness. To further counteract potential value overestimation or underestimation caused by the RCB operator, we introduce two techniques, the dynamic value penalty and the Huber loss, into our framework, resulting in the practical \textbf{D}ual-\textbf{RO}bust \textbf{C}ross-domain \textbf{O}ffline RL (DROCO) algorithm. Extensive empirical results across various dynamics shift scenarios show that DROCO outperforms strong baselines and exhibits enhanced robustness to dynamics perturbations.

Paperid: 826, https://arxiv.org/pdf/2512.02435.pdf

Abstract:
Cross-Domain Offline Reinforcement Learning aims to train an agent deployed in the target environment, leveraging both a limited target domain dataset and a source domain dataset with (possibly) sufficient data coverage. Due to the underlying dynamics misalignment between the source and target domain, simply merging the data from two datasets may incur inferior performance. Recent advances address this issue by selectively sharing source domain samples that exhibit dynamics alignment with the target domain. However, these approaches focus solely on dynamics alignment and overlook \textit{value alignment}, i.e., selecting high-quality, high-value samples from the source domain. In this paper, we first demonstrate that both dynamics alignment and value alignment are essential for policy learning, by examining the limitations of the current theoretical framework for cross-domain RL and establishing a concrete sub-optimality gap of a policy trained on the source domain and evaluated on the target domain. Motivated by the theoretical insights, we propose to selectively share those source domain samples with both high dynamics and value alignment and present our \textbf{\underline{D}}ynamics- and \textbf{\underline{V}}alue-aligned \textbf{\underline{D}}ata \textbf{\underline{F}}iltering (DVDF) method. We design a range of dynamics shift settings, including kinematic and morphology shifts, and evaluate DVDF on various tasks and datasets, as well as in challenging extremely low-data settings where the target domain dataset contains only 5,000 transitions. Extensive experiments demonstrate that DVDF consistently outperforms prior strong baselines and delivers exceptional performance across multiple tasks and datasets.

Paperid: 827, https://arxiv.org/pdf/2512.02353.pdf

Abstract:
Orthogonal time frequency space (OTFS) modulation has been proposed to meet the demand for reliable communication in high-mobility scenarios for future wireless networks. However, in multi-user OTFS systems, conventional embedded pilot schemes require independent pilot allocation for each user, leading to linearly increasing pilot overhead. To address these issues, in this paper, we investigate the uplink channel estimation and pilot design for multi-user multiple-input multiple-output (MIMO)-OTFS systems. We propose a multi-dimensional decomposition-based channel estimation algorithm. Specifically, the proposed algorithm first estimates the angles of arrivals (AoAs) via subspace decomposition-based method. A spatial projection matrix, constructed from the estimated AOAs, decouples the received signal by propagation path subspace, effectively mitigating inter-path interference. The remaining fractional delay and Doppler can be obtained by a compressed sensing (CS)-based off-grid channel estimation method. Furthermore, to reduce the pilot overhead in multi-user OTFS systems, this paper proposes a novel cyclic shift embedded pilot (CSEP) structure, which can reuse users through cyclic shift-orthogonality of Zadoff-Chu (ZC) sequences. Compared with conventional embedded pilot structures, the CSEP structure can save over 30\% of pilot overhead. Finally, an imporved channel estimation method based on the CSEP structure is proposed. Simulation results demonstrate that it achieves superior performance in channel estimation. Moreover, the proposed CSEP structure and channel estimation algorithm achieve a favorable balance between computational complexity, estimation accuracy, and bit error rate (BER) performance.

Paperid: 828, https://arxiv.org/pdf/2512.02032.pdf

Abstract:
Structural connectomes are detailed graphs that map how different brain regions are physically connected, offering critical insight into aging, cognition, and neurodegenerative diseases. However, these connectomes are high-dimensional and densely interconnected, which makes them difficult to interpret and analyze at scale. While low-dimensional spaces like PCA and autoencoders are often used to capture major sources of variation, their latent spaces are generally continuous and cannot fully reflect the mixed nature of variability in connectomes, which include both continuous (e.g., connectivity strength) and discrete factors (e.g., imaging site). Motivated by this, we propose a variational autoencoder (VAE) with a hybrid latent space that jointly models the discrete and continuous components. We analyze a large dataset of 5,761 connectomes from six Alzheimer's disease studies with ten acquisition protocols. Each connectome represents a single scan from a unique subject (3579 females, 2182 males), aged 22 to 102, with 4338 cognitively normal, 809 with mild cognitive impairment (MCI), and 614 with Alzheimer's disease (AD). Each connectome contains 121 brain regions defined by the BrainCOLOR atlas. We train our hybrid VAE in an unsupervised way and characterize what each latent component captures. We find that the discrete space is particularly effective at capturing subtle site-related differences, achieving an Adjusted Rand Index (ARI) of 0.65 with site labels, significantly outperforming PCA and a standard VAE followed by clustering (p < 0.05). These results demonstrate that the hybrid latent space can disentangle distinct sources of variability in connectomes in an unsupervised manner, offering potential for large-scale connectome analysis.

Paperid: 829, https://arxiv.org/pdf/2512.01302.pdf

Abstract:
Despite recent text-to-image models achieving highfidelity text rendering, they still struggle with long or multiple texts due to diluted global attention. We propose DCText, a training-free visual text generation method that adopts a divide-and-conquer strategy, leveraging the reliable short-text generation of Multi-Modal Diffusion Transformers. Our method first decomposes a prompt by extracting and dividing the target text, then assigns each to a designated region. To accurately render each segment within their regions while preserving overall image coherence, we introduce two attention masks - Text-Focus and Context-Expansion - applied sequentially during denoising. Additionally, Localized Noise Initialization further improves text accuracy and region alignment without increasing computational cost. Extensive experiments on single- and multisentence benchmarks show that DCText achieves the best text accuracy without compromising image quality while also delivering the lowest generation latency.

Paperid: 830, https://arxiv.org/pdf/2512.00834.pdf

Abstract:
Efficient information exchange and reliable contextual reasoning are essential for vehicle-to-everything (V2X) networks. Conventional communication schemes often incur significant transmission overhead and latency, while existing trajectory prediction models generally lack environmental perception and logical inference capabilities. This paper presents a trajectory prediction framework that integrates semantic communication with Agentic AI to enhance predictive performance in vehicular environments. In vehicle-to-infrastructure (V2I) communication, a feature-extraction agent at the Roadside Unit (RSU) derives compact representations from historical vehicle trajectories, followed by semantic reasoning performed by a semantic-analysis agent. The RSU then transmits both feature representations and semantic insights to the target vehicle via semantic communication, enabling the vehicle to predict future trajectories by combining received semantics with its own historical data. In vehicle-to-vehicle (V2V) communication, each vehicle performs local feature extraction and semantic analysis while receiving predicted trajectories from neighboring vehicles, and jointly utilizes this information for its own trajectory prediction. Extensive experiments across diverse communication conditions demonstrate that the proposed method significantly outperforms baseline schemes, achieving up to a 47.5% improvement in prediction accuracy under low signal-to-noise ratio (SNR) conditions.

Paperid: 831, https://arxiv.org/pdf/2512.00491.pdf

Abstract:
The Transmission Control Protocol (TCP) relies on a state machine and deterministic arithmetic to ensure reliable connections. However, traditional protocol logic driven by hard-coded state machines struggles to meet the demands of intelligent and autonomous network architectures. Here, we adopt the agentic AI-based paradigm, driven by Large Language Models (LLMs), characterized by context perception, autonomous reasoning, and tool use. Based on this, we propose Smart-TCP, which re-imagines TCP's core control logic as an autonomous agent. Specifically, the proposed architecture employs a context aggregation mechanism to synthesize the protocol context, utilizes the LLM for autonomous logical reasoning, and invokes an Arithmetic Logic Unit (ALU) as a tool for computation. Furthermore, we establish a dual-agent interaction framework based on this architecture and implement TCP protocol interactions. Experiments demonstrate that the Smart-TCP agent excels in static prediction and error detection, achieving a 93.33% success rate in end-to-end sessions. These results strongly validate the technical feasibility of an agentic AI-based TCP protocol.

Paperid: 832, https://arxiv.org/pdf/2512.00490.pdf

Abstract:
Vehicle-to-Everything (V2X) cooperation is reshaping traffic safety from an ego-centric sensing problem into one of collective intelligence. This survey structures recent progress within a unified Sensor-Perception-Decision (SPD) framework that formalizes how safety emerges from the interaction of distributed sensing, cooperative perception, and coordinated decision-making across vehicles and infrastructure. Rather than centering on link protocols or message formats, we focus on how shared evidence, predictive reasoning, and human-aligned interventions jointly enable proactive risk mitigation. Within this SPD lens, we synthesize advances in cooperative perception, multi-modal forecasting, and risk-aware planning, emphasizing how cross-layer coupling turns isolated detections into calibrated, actionable understanding. Timing, trust, and human factors are identified as cross-cutting constraints that determine whether predictive insights are delivered early enough, with reliable confidence, and in forms that humans and automated controllers can use. Compared with prior V2X safety surveys, this work (i) organizes the literature around a formal SPD safety loop and (ii) systematically analyzes research evolution and evaluation gaps through a PRISMA-guided bibliometric study of hundreds of publications from 2016-2025. The survey concludes with a roadmap toward cooperative safety intelligence, outlining SPD-based design principles and evaluation practices for next-generation V2X safety systems.

Paperid: 833, https://arxiv.org/pdf/2512.07351.pdf

Abstract:
The increasing use of synthetic media, particularly deepfakes, is an emerging challenge for digital content verification. Although recent studies use both audio and visual information, most integrate these cues within a single model, which remains vulnerable to modality mismatches, noise, and manipulation. To address this gap, we propose DeepAgent, an advanced multi-agent collaboration framework that simultaneously incorporates both visual and audio modalities for the effective detection of deepfakes. DeepAgent consists of two complementary agents. Agent-1 examines each video with a streamlined AlexNet-based CNN to identify the symbols of deepfake manipulation, while Agent-2 detects audio-visual inconsistencies by combining acoustic features, audio transcriptions from Whisper, and frame-reading sequences of images through EasyOCR. Their decisions are fused through a Random Forest meta-classifier that improves final performance by taking advantage of the different decision boundaries learned by each agent. This study evaluates the proposed framework using three benchmark datasets to demonstrate both component-level and fused performance. Agent-1 achieves a test accuracy of 94.35% on the combined Celeb-DF and FakeAVCeleb datasets. On the FakeAVCeleb dataset, Agent-2 and the final meta-classifier attain accuracies of 93.69% and 81.56%, respectively. In addition, cross-dataset validation on DeepFakeTIMIT confirms the robustness of the meta-classifier, which achieves a final accuracy of 97.49%, and indicates a strong capability across diverse datasets. These findings confirm that hierarchy-based fusion enhances robustness by mitigating the weaknesses of individual modalities and demonstrate the effectiveness of a multi-agent approach in addressing diverse types of manipulations in deepfakes.

Paperid: 834, https://arxiv.org/pdf/2512.06906.pdf

Abstract:
Detecting the anomalies of web applications, important infrastructures for running modern companies and governments, is crucial for providing reliable web services. Many modern web applications operate on web APIs (e.g., RESTful, SOAP, and WebSockets), their exposure invites intended attacks or unintended illegal visits, causing abnormal system behaviors. However, such anomalies can share very similar logs with normal logs, missing crucial information (which could be in database) for log discrimination. Further, log instances can be also noisy, which can further mislead the state-of-the-art log learning solutions to learn spurious correlation, resulting superficial models and rules for anomaly detection. In this work, we propose MINES which infers explainable API invariants for anomaly detection from the schema level instead of detailed raw log instances, which can (1) significantly discriminate noise in logs to identify precise normalities and (2) detect abnormal behaviors beyond the instrumented logs. Technically, MINES (1) converts API signatures into table schema to enhance the original database shema; and (2) infers the potential database constraints on the enhanced database schema to capture the potential relationships between APIs and database tables. MINES uses LLM for extracting potential relationship based on two given table structures; and use normal log instances to reject and accept LLM-generated invariants. Finally, MINES translates the inferred constraints into invariants to generate Python code for verifying the runtime logs. We extensively evaluate MINES on web-tamper attacks on the benchmarks of TrainTicket, NiceFish, Gitea, Mastodon, and NextCloud against baselines such as LogRobust, LogFormer, and WebNorm. The results show that MINES achieves high recall for the anomalies while introducing almost zero false positives, indicating a new state-of-the-art.

Paperid: 835, https://arxiv.org/pdf/2512.06363.pdf

Abstract:
Real-world face recognition systems are vulnerable to both physical presentation attacks (PAs) and digital forgery attacks (DFs). We aim to achieve comprehensive protection of biometric data by implementing a unified physical-digital defense framework with advanced detection. Existing approaches primarily employ CLIP with regularization constraints to enhance model generalization across both tasks. However, these methods suffer from conflicting optimization directions between physical and digital attack detection under same category prompt spaces. To overcome this limitation, we propose a Spoofing-aware Prompt Learning for Unified Attack Detection (SPL-UAD) framework, which decouples optimization branches for physical and digital attacks in the prompt space. Specifically, we construct a learnable parallel prompt branch enhanced with adaptive Spoofing Context Prompt Generation, enabling independent control of optimization for each attack type. Furthermore, we design a Cues-awareness Augmentation that leverages the dual-prompt mechanism to generate challenging sample mining tasks on data, significantly enhancing the model's robustness against unseen attack types. Extensive experiments on the large-scale UniAttackDataPlus dataset demonstrate that the proposed method achieves significant performance improvements in unified attack detection tasks.

Paperid: 836, https://arxiv.org/pdf/2512.05959.pdf

Abstract:
Vision-language models (VLMs) have achieved strong performance in visual question answering (VQA), yet they remain constrained by static training data. Retrieval-Augmented Generation (RAG) mitigates this limitation by enabling access to up-to-date, culturally grounded, and multilingual information; however, multilingual multimodal RAG remains largely underexplored. We introduce M4-RAG, a massive-scale benchmark covering 42 languages and 56 regional dialects and registers, comprising over 80,000 culturally diverse image-question pairs for evaluating retrieval-augmented VQA across languages and modalities. To balance realism with reproducibility, we build a controlled retrieval environment containing millions of carefully curated multilingual documents relevant to the query domains, approximating real-world retrieval conditions while ensuring consistent experimentation. Our systematic evaluation reveals that although RAG consistently benefits smaller VLMs, it fails to scale to larger models and often even degrades their performance, exposing a critical mismatch between model size and current retrieval effectiveness. M4-RAG provides a foundation for advancing next-generation RAG systems capable of reasoning seamlessly across languages, modalities, and cultural contexts.

Paperid: 837, https://arxiv.org/pdf/2512.05920.pdf

Abstract:
Orthognathic surgery is a crucial intervention for correcting dentofacial skeletal deformities to enhance occlusal functionality and facial aesthetics. Accurate postoperative facial appearance prediction remains challenging due to the complex nonlinear interactions between skeletal movements and facial soft tissue. Existing biomechanical, parametric models and deep-learning approaches either lack computational efficiency or fail to fully capture these intricate interactions. To address these limitations, we propose Neural Implicit Craniofacial Model (NICE) which employs implicit neural representations for accurate anatomical reconstruction and surgical outcome prediction. NICE comprises a shape module, which employs region-specific implicit Signed Distance Function (SDF) decoders to reconstruct the facial surface, maxilla, and mandible, and a surgery module, which employs region-specific deformation decoders. These deformation decoders are driven by a shared surgical latent code to effectively model the complex, nonlinear biomechanical response of the facial surface to skeletal movements, incorporating anatomical prior knowledge. The deformation decoders output point-wise displacement fields, enabling precise modeling of surgical outcomes. Extensive experiments demonstrate that NICE outperforms current state-of-the-art methods, notably improving prediction accuracy in critical facial regions such as lips and chin, while robustly preserving anatomical integrity. This work provides a clinically viable tool for enhanced surgical planning and patient consultation in orthognathic procedures.

Paperid: 838, https://arxiv.org/pdf/2512.05551.pdf

Abstract:
Code ownership is central to ensuring accountability and maintaining quality in large-scale software development. Yet, as external threats such as software supply chain attacks on project health and quality assurance increase, mechanisms for assigning and enforcing responsibility have become increasingly critical. In 2017, GitHub introduced the CODEOWNERS feature, which automatically designates reviewers for specific files to strengthen accountability and protect critical parts of the codebase. Despite its potential, little is known about how CODEOWNERS is actually adopted and practiced. We present the first large-scale empirical study of CODEOWNERS usage across over 844,000 pull requests with 1.9 million comments and over 2 million reviews. We identify 10,287 code owners to track their review activities. Results indicate that codeowners tend to adhere the rules specified in the CODEOWNERS file, exhibit similar collaborative behaviours to traditional metrics of ownership, but tend to contribute to a smoother and faster PR workflow over time. Finally, using regression discontinuity design (RDD) analysis, we find that repositories adopting CODEOWNERS experience shifts in review dynamics, as ownership redistributes review responsibilities away from core developers. Our results position CODEOWNERS as a promising yet underutilized mechanism for improving software governance and resilience. We discuss how projects can leverage this alternative ownership method as a perspective to enhance security, accountability, and workflow efficiency in open-source development.

Paperid: 839, https://arxiv.org/pdf/2512.05121.pdf

Abstract:
PESTalk is a novel method for generating 3D facial animations with personalized emotional styles directly from speech. It overcomes key limitations of existing approaches by introducing a Dual-Stream Emotion Extractor (DSEE) that captures both time and frequency-domain audio features for fine-grained emotion analysis, and an Emotional Style Modeling Module (ESMM) that models individual expression patterns based on voiceprint characteristics. To address data scarcity, the method leverages a newly constructed 3D-EmoStyle dataset. Evaluations demonstrate that PESTalk outperforms state-of-the-art methods in producing realistic and personalized facial animations.

Paperid: 840, https://arxiv.org/pdf/2512.04220.pdf

Abstract:
Tool-integrated (TI) reinforcement learning (RL) enables large language models (LLMs) to perform multi-step reasoning by interacting with external tools such as search engines and retrievers. Group Relative Policy Optimization (GRPO), exemplified by the recent Search-R1, offers fast convergence and a value-free formulation that makes it appealing for this setting, yet consistently suffers from training collapse. We identify Lazy Likelihood Displacement (LLD), a systematic reduction or stagnation in the likelihood of both correct and incorrect responses, as the core mechanism driving this failure. LLD emerges early and triggers a self-reinforcing LLD Death Spiral, where declining likelihood leads to low-confidence responses, inflating gradients, and ultimately causing collapse. We empirically characterize this process across models on a Search-R1-style, search-integrated question answering task, revealing a consistent three-phase trajectory: early stagnation, steady decay, and accelerated collapse. To address this, we propose a lightweight likelihood-preserving regularization LLDS for GRPO that activates only when a trajectory's likelihood decreases, and regularizes only the tokens responsible. This fine-grained structure mitigates LLD with minimal interference to optimization. Across seven open-domain and multi-hop QA benchmarks, our method stabilizes training, prevents gradient explosion, and yields substantial performance improvements, including +37.8% gains on Qwen2.5-3B and +32.0% gains on Qwen2.5-7B. Our results establish LLD as a fundamental bottleneck in GRPO-based TIRL and provide a practical path toward stable, scalable training of tool-integrated LLM.

Paperid: 841, https://arxiv.org/pdf/2512.03238.pdf

Abstract:
High quality data is needed to unlock the full potential of AI for end users. However finding new sources of such data is getting harder: most publicly-available human generated data will soon have been used. Additionally, publicly available data often is not representative of users of a particular system -- for example, a research speech dataset of contractors interacting with an AI assistant will likely be more homogeneous, well articulated and self-censored than real world commands that end users will issue. Therefore unlocking high-quality data grounded in real user interactions is of vital interest. However, the direct use of user data comes with significant privacy risks. Differential Privacy (DP) is a well established framework for reasoning about and limiting information leakage, and is a gold standard for protecting user privacy. The focus of this work, \emph{Differentially Private Synthetic data}, refers to synthetic data that preserves the overall trends of source data,, while providing strong privacy guarantees to individuals that contributed to the source dataset. DP synthetic data can unlock the value of datasets that have previously been inaccessible due to privacy concerns and can replace the use of sensitive datasets that previously have only had rudimentary protections like ad-hoc rule-based anonymization. In this paper we explore the full suite of techniques surrounding DP synthetic data, the types of privacy protections they offer and the state-of-the-art for various modalities (image, tabular, text and decentralized). We outline all the components needed in a system that generates DP synthetic data, from sensitive data handling and preparation, to tracking the use and empirical privacy testing. We hope that work will result in increased adoption of DP synthetic data, spur additional research and increase trust in DP synthetic data approaches.

Paperid: 842, https://arxiv.org/pdf/2512.03086.pdf

Abstract:
Large language models (LLMs) have shown remarkable capabilities in code translation, yet their performance deteriorates in low-resource programming domains such as Fortran and emerging frameworks like CUDA, where high-quality parallel data are scarce. We present an automated dataset generation pipeline featuring a dual-LLM Questioner-Solver design that incorporates external knowledge from compilers and runtime feedback. Beyond traditional source-target code pair datasets, our approach additionally generates (1) verified translations with unit tests for assessing functional consistency, and (2) multi-turn dialogues that capture the reasoning process behind translation refinement. Applied to Fortran -> C++ and C++ -> CUDA, the pipeline yields 3.64k and 3.93k dialogues, respectively. Fine-tuning on this data yields dramatic improvements in functional correctness, boosting unit test success rates by over 56% on the challenging C++-to-CUDA task. We show this data enables a 7B open-weight model to significantly outperform larger proprietary systems on key metrics like compilation success.

Paperid: 843, https://arxiv.org/pdf/2512.02932.pdf

Abstract:
Novel view synthesis (NVS) is crucial in computer vision and graphics, with wide applications in AR, VR, and autonomous driving. While 3D Gaussian Splatting (3DGS) enables real-time rendering with high appearance fidelity, it suffers from multi-view inconsistencies, limiting geometric accuracy. In contrast, 2D Gaussian Splatting (2DGS) enforces multi-view consistency but compromises texture details. To address these limitations, we propose Exchangeable Gaussian Splatting (EGGS), a hybrid representation that integrates 2D and 3D Gaussians to balance appearance and geometry. To achieve this, we introduce Hybrid Gaussian Rasterization for unified rendering, Adaptive Type Exchange for dynamic adaptation between 2D and 3D Gaussians, and Frequency-Decoupled Optimization that effectively exploits the strengths of each type of Gaussian representation. Our CUDA-accelerated implementation ensures efficient training and inference. Extensive experiments demonstrate that EGGS outperforms existing methods in rendering quality, geometric accuracy, and efficiency, providing a practical solution for high-quality NVS.

Paperid: 844, https://arxiv.org/pdf/2512.02882.pdf

Abstract:
Test-time policy optimization enables large language models (LLMs) to adapt to distribution shifts by leveraging feedback from self-generated rollouts. However, existing methods rely on fixed-budget majority voting to estimate rewards, incurring substantial computational redundancy. We propose Optimal Rollout Allocation for Test-time Policy Optimization (OptPO), a principled framework that adaptively allocates inference budgets. By formulating the voting process as a Bayesian sequential probability ratio test, OptPO dynamically halts sampling once the posterior confidence in a consensus answer exceeds a specified threshold. Crucially, it utilizes the retained rollouts for on-policy updates, seamlessly integrating with algorithms like PPO or GRPO without requiring ground-truth labels. Across diverse reasoning benchmarks, OptPO significantly reduces rollout overhead compared to fixed-sample baselines while preserving or improving accuracy. By unifying statistically optimal stopping with test-time learning, OptPO offers a computationally efficient paradigm for test-time adaptation. The source code will be open upon acceptance at https://open-upon-acceptance.

Paperid: 845, https://arxiv.org/pdf/2512.02566.pdf

Abstract:
There is a growing interest in developing strong biomedical vision-language models. A popular approach to achieve robust representations is to use web-scale scientific data. However, current biomedical vision-language pretraining typically compresses rich scientific figures and text into coarse figure-level pairs, discarding the fine-grained correspondences that clinicians actually rely on when zooming into local structures. To tackle this issue, we introduce Panel2Patch, a novel data pipeline that mines hierarchical structure from existing biomedical scientific literature, i.e., multi-panel, marker-heavy figures and their surrounding text, and converts them into multi-granular supervision. Given scientific figures and captions, Panel2Patch parses layouts, panels, and visual markers, then constructs hierarchical aligned vision-language pairs at the figure, panel, and patch levels, preserving local semantics instead of treating each figure as a single data sample. Built on this hierarchical corpus, we develop a granularity-aware pretraining strategy that unifies heterogeneous objectives from coarse didactic descriptions to fine region-focused phrases. By applying Panel2Patch to only a small set of the literature figures, we extract far more effective supervision than prior pipelines, enabling substantially better performance with less pretraining data.

Paperid: 846, https://arxiv.org/pdf/2512.02541.pdf

Abstract:
Since DUSt3R, models such as VGGT and $π^3$ have shown strong multi-view 3D performance, but their heavy reliance on global self-attention results in high computational cost. Existing sparse-attention variants offer partial speedups, yet lack a systematic analysis of how global attention contributes to multi-view reasoning. In this paper, we first conduct an in-depth investigation of the global attention modules in VGGT and $π^3$ to better understand their roles. Our analysis reveals a clear division of roles in the alternating global-frame architecture: early global layers do not form meaningful correspondences, middle layers perform cross-view alignment, and last layers provide only minor refinements. Guided by these findings, we propose a training-free two-step acceleration scheme: (1) converting early global layers into frame attention, and (2) subsampling global attention by subsampling K/V over patch tokens with diagonal preservation and a mean-fill component. We instantiate this strategy on VGGT and $π^3$ and evaluate across standard pose and point-map benchmarks. Our method achieves up to $8$-$10\times$ speedup in inference time while matching or slightly improving the accuracy of the original models, and remains robust even in extremely dense multi-view settings where prior sparse-attention baselines fail.

Paperid: 847, https://arxiv.org/pdf/2512.02461.pdf

Abstract:
With the evolution of wireless systems toward large-scale arrays and high-frequency reconfigurable architectures, fluid antenna systems (FAS) operating in the near-field (NF) regime provide new degrees of freedom (DoF) for physical layer security (PLS). This paper proposes an artificial-noise (AN)-aided PLS scheme for NF fluid-antenna multiple-input multiple-output (FA-MIMO) systems, with joint beamforming (BF) and AN design for both compact and large arrays. An alternating-optimization (AO) framework addresses the sparsity-constrained non-convex design by splitting it into a continuous BF/AN joint-design subproblem and a discrete FAS port-selection subproblem. Closed-form fully digital BF/AN solutions are obtained via a generalized spectral water-filling procedure within a block coordinate descent (BCD) surrogate and realized by a hardware-efficient hybrid beamforming (HBF) architecture that embeds AN in the baseband without extra radio-frequency (RF) chains. For FAS port selection, a row-energy based prune--refit rule, aligned with Karush--Kuhn--Tucker (KKT) conditions of a group-sparsity surrogate, enables efficient active-port determination. Simulation results confirm that the proposed design exploits the geometry and position-domain DoF of FAS and significantly improves secrecy performance, particularly for non-extremely-large arrays where NF beam focusing alone is inadequate.

Paperid: 848, https://arxiv.org/pdf/2512.02381.pdf

Abstract:
The electrification of off-road heavy equipment presents operational challenges for agencies serving remote sites with limited fixed charging infrastructure. Existing mobile fast charging vehicle (MFCV) planning approaches typically treat fleet design and routing as separate problems, fixing vehicle characteristics before dispatch. This paper formulates a fleet size and mix capacitated vehicle routing problem with time windows (FSMCVRPTW) for MFCV deployment, jointly optimizing fleet composition, charger specifications, routing, and scheduling within a unified mixed-integer linear program. The model incorporates heterogeneous MFCV types with varying power ratings, battery capacities, fuel range, and cost structures, minimizing total daily cost from labor, fuel, amortized capital expenditure, and energy purchase under temporal service windows, resource budgets, and energy-delivery constraints. The formulation is implemented in Python/Gurobi and applied to two case studies using California Department of Transportation wheel-loader data in Los Angeles (dense urban) and Truckee (sparse mountainous). Results show that simultaneous optimization yields compact, well-utilized fleets that meet all service windows while revealing strong sensitivity of unit cost to demand density and geography. The proposed FSMCVRPTW framework provides a generalizable decision-support methodology that co-designs fleet size, charger power, routing, and service schedules in a single optimization layer for context-aware, cost-efficient mobile fast charging.

Paperid: 849, https://arxiv.org/pdf/2512.01821.pdf

Abstract:
Spatial reasoning, the ability to understand and interpret the 3D structure of the world, is a critical yet underdeveloped capability in Multimodal Large Language Models (MLLMs). Current methods predominantly rely on verbal descriptive tuning, which suffers from visual illiteracy, i.e., they learn spatial concepts through textual symbols alone, devoid of connection to their visual manifestations. To bridge this gap, this paper introduces MILO, an Implicit spatIaL wOrld modeling paradigm that simulates human-like spatial imagination. MILO integrates a visual generator to provide geometry-aware feedback, thereby implicitly grounding the MLLM's symbolic reasoning in perceptual experience. Complementing this paradigm, we propose RePE (Relative Positional Encoding), a novel encoding scheme that captures relative camera-pose transformations, offering superior performance over absolute coordinate systems. To support the training, we construct GeoGen, a large-scale Geometry-aware Generative dataset with approximately 2,241 videos and 67,827 observation-action-outcome triplets. Experiments demonstrate that our approach significantly enhances spatial reasoning capabilities across multiple baselines and benchmarks, offering a more holistic understanding of 3D space.

Paperid: 850, https://arxiv.org/pdf/2512.01286.pdf

Abstract:
Flow matching has recently emerged as a promising alternative to diffusion-based generative models, offering faster sampling and simpler training by learning continuous flows governed by ordinary differential equations. Despite growing empirical success, the theoretical understanding of flow matching remains limited, particularly in terms of sample complexity results. In this work, we provide the first analysis of the sample complexity for flow-matching based generative models without assuming access to the empirical risk minimizer (ERM) of the loss function for estimating the velocity field. Under standard assumptions on the loss function for velocity field estimation and boundedness of the data distribution, we show that a sufficiently expressive neural network can learn a velocity field such that with $\mathcal{O}(ε^{-4})$ samples, such that the Wasserstein-2 distance between the learned and the true distribution is less than $\mathcal{O}(ε)$. The key technical idea is to decompose the velocity field estimation error into neural-network approximation error, statistical error due to the finite sample size, and optimization error due to the finite number of optimization steps for estimating the velocity field. Each of these terms are then handled via techniques that may be of independent interest.

Paperid: 851, https://arxiv.org/pdf/2512.00939.pdf

Abstract:
Recent progress in contact-rich robotic manipulation has been striking, yet most deployed systems remain confined to simple, scripted routines. One of the key barriers is the lack of motion planning algorithms that can provide verifiable guarantees for safety, efficiency and reliability. To address this, a family of algorithms called Constant-Time Motion Planning (CTMP) was introduced, which leverages a preprocessing phase to enable collision-free motion queries in a fixed, user-specified time budget (e.g., 10 milliseconds). However, existing CTMP methods do not explicitly incorporate the manipulation behaviors essential for object handling. To bridge this gap, we introduce the \textit{Behavioral Constant-Time Motion Planner} (B-CTMP), an algorithm that extends CTMP to solve a broad class of two-step manipulation tasks: (1) a collision-free motion to a behavior initiation state, followed by (2) execution of a manipulation behavior (such as grasping or insertion) to reach the goal. By precomputing compact data structures, B-CTMP guarantees constant-time query in mere milliseconds while ensuring completeness and successful task execution over a specified set of states. We evaluate B-CTMP on two canonical manipulation tasks in simulation, shelf picking and plug insertion,and demonstrate its effectiveness on a real robot. Our results show that B-CTMP unifies collision-free planning and object manipulation within a single constant-time framework, providing provable guarantees of speed and success for manipulation in semi-structured environments.

Paperid: 852, https://arxiv.org/pdf/2512.00632.pdf

Abstract:
Perfect $L_p$ sampling in a stream was introduced by Jayaram and Woodruff (FOCS 2018) as a streaming primitive which, given turnstile updates to a vector $x \in \{-\text{poly}(n), \ldots, \text{poly}(n)\}^n$, outputs an index $i^* \in \{1, 2, \ldots, n\}$ such that the probability of returning index $i$ is exactly \[\Pr[i^* = i] = \frac{|x_i|^p}{\|x\|_p^p} \pm \frac{1}{n^C},\] where $C > 0$ is an arbitrarily large constant. Jayaram and Woodruff achieved the optimal $\tilde{O}(\log^2 n)$ bits of memory for $0 < p < 2$, but their update time is at least $n^C$ per stream update. Thus an important open question is to achieve efficient update time while maintaining optimal space. For $0 < p < 2$, we give the first perfect $L_p$-sampler with the same optimal amount of memory but with only $\text{poly}(\log n)$ update time. Crucial to our result is an efficient simulation of a sum of reciprocals of powers of truncated exponential random variables by approximating its characteristic function, using the Gil-Pelaez inversion formula, and applying variants of the trapezoid formula to quickly approximate it.

Paperid: 853, https://arxiv.org/pdf/2512.00290.pdf

Abstract:
Large language models (LLMs) demonstrate significant potential for educational applications. However, their unscrutinized deployment poses risks to educational standards, underscoring the need for rigorous evaluation. We introduce EduEval, a comprehensive hierarchical benchmark for evaluating LLMs in Chinese K-12 education. This benchmark makes three key contributions: (1) Cognitive Framework: We propose the EduAbility Taxonomy, which unifies Bloom's Taxonomy and Webb's Depth of Knowledge to organize tasks across six cognitive dimensions including Memorization, Understanding, Application, Reasoning, Creativity, and Ethics. (2) Authenticity: Our benchmark integrates real exam questions, classroom conversation, student essays, and expert-designed prompts to reflect genuine educational challenges; (3) Scale: EduEval comprises 24 distinct task types with over 11,000 questions spanning primary to high school levels. We evaluate 14 leading LLMs under both zero-shot and few-shot settings, revealing that while models perform well on factual tasks, they struggle with classroom dialogue classification and exhibit inconsistent results in creative content generation. Interestingly, several open source models outperform proprietary systems on complex educational reasoning. Few-shot prompting shows varying effectiveness across cognitive dimensions, suggesting that different educational objectives require tailored approaches. These findings provide targeted benchmarking metrics for developing LLMs specifically optimized for diverse Chinese educational tasks.

Paperid: 854, https://arxiv.org/pdf/2512.00041.pdf

Abstract:
Vision-and-Language Navigation (VLN) requires agents to follow language instructions while acting in continuous real-world spaces. Prior image imagination based VLN work shows benefits for discrete panoramas but lacks online, action-conditioned predictions and does not produce explicit planning values; moreover, many methods replace the planner with long-horizon objectives that are brittle and slow. To bridge this gap, we propose VISTAv2, a generative world model that rolls out egocentric future views conditioned on past observations, candidate action sequences, and instructions, and projects them into an online value map for planning. Unlike prior approaches, VISTAv2 does not replace the planner. The online value map is fused at score level with the base objective, providing reachability and risk-aware guidance. Concretely, we employ an action-aware Conditional Diffusion Transformer video predictor to synthesize short-horizon futures, align them with the natural language instruction via a vision-language scorer, and fuse multiple rollouts in a differentiable imagination-to-value head to output an imagined egocentric value map. For efficiency, rollouts occur in VAE latent space with a distilled sampler and sparse decoding, enabling inference on a single consumer GPU. Evaluated on MP3D and RoboTHOR, VISTAv2 improves over strong baselines, and ablations show that action-conditioned imagination, instruction-guided value fusion, and the online value-map planner are all critical, suggesting that VISTAv2 offers a practical and interpretable route to robust VLN.

Paperid: 855, https://arxiv.org/pdf/2512.07799.pdf

Abstract:
Carbon-aware schedulers aim to reduce the operational carbon footprint of data centers by running flexible workloads during periods of low carbon intensity. Most schedulers treat workloads as single monolithic tasks, ignoring that many jobs, like video encoding or offline inference, consist of smaller tasks with specific dependencies and resource needs; however, knowledge of this structure enables opportunities for greater carbon efficiency. We quantify the maximum benefit of a dependency-aware approach for batch workloads. We model the problem as a flexible job-shop scheduling variant and use an offline solver to compute upper bounds on carbon and energy savings. Results show up to $25\%$ lower carbon emissions on average without increasing the optimal makespan (total job completion time) compared to a makespan-only baseline. Although in heterogeneous server setup, these schedules may use more energy than energy-optimal ones. Our results also show that allowing twice the optimal makespan nearly doubles the carbon savings, underscoring the tension between carbon, energy, and makespan. We also highlight key factors such as job structure and server count influence the achievable carbon reductions.

Paperid: 856, https://arxiv.org/pdf/2512.07733.pdf

Abstract:
Despite advancements in Multi-modal Large Language Models (MLLMs) for scene understanding, their performance on complex spatial reasoning tasks requiring mental simulation remains significantly limited. Current methods often rely on passive observation of spatial data, failing to internalize an active mental imagery process. To bridge this gap, we propose SpatialDreamer, a reinforcement learning framework that enables spatial reasoning through a closedloop process of active exploration, visual imagination via a world model, and evidence-grounded reasoning. To address the lack of fine-grained reward supervision in longhorizontal reasoning tasks, we propose Geometric Policy Optimization (GeoPO), which introduces tree-structured sampling and step-level reward estimation with geometric consistency constraints. Extensive experiments demonstrate that SpatialDreamer delivers highly competitive results across multiple challenging benchmarks, signifying a critical advancement in human-like active spatial mental simulation for MLLMs.

Paperid: 857, https://arxiv.org/pdf/2512.07459.pdf

Abstract:
Generating realistic human geometry animations remains a challenging task, as it requires modeling natural clothing dynamics with fine-grained geometric details under limited data. To address these challenges, we propose two novel designs. First, we propose a compact distribution-based latent representation that enables efficient and high-quality geometry generation. We improve upon previous work by establishing a more uniform mapping between SMPL and avatar geometries. Second, we introduce a generative animation model that fully exploits the diversity of limited motion data. We focus on short-term transitions while maintaining long-term consistency through an identity-conditioned design. These two designs formulate our method as a two-stage framework: the first stage learns a latent space, while the second learns to generate animations within this latent space. We conducted experiments on both our latent space and animation model. We demonstrate that our latent space produces high-fidelity human geometry surpassing previous methods ($90\%$ lower Chamfer Dist.). The animation model synthesizes diverse animations with detailed and natural dynamics ($2.2 \times$ higher user study score), achieving the best results across all evaluation metrics.

Paperid: 858, https://arxiv.org/pdf/2512.07384.pdf

Abstract:
In recommender systems, user-item interactions can be modeled as a bipartite graph, where user and item nodes are connected by undirected edges. This graph-based view has motivated the rapid adoption of graph neural networks (GNNs), which often outperform collaborative filtering (CF) methods such as latent factor models, deep neural networks, and generative strategies. Yet, despite their empirical success, the reasons why GNNs offer systematic advantages over other CF approaches remain only partially understood. This monograph advances a topology-centered perspective on GNN-based recommendation. We argue that a comprehensive understanding of these models' performance should consider the structural properties of user-item graphs and their interaction with GNN architectural design. To support this view, we introduce a formal taxonomy that distills common modeling patterns across eleven representative GNN-based recommendation approaches and consolidates them into a unified conceptual pipeline. We further formalize thirteen classical and topological characteristics of recommendation datasets and reinterpret them through the lens of graph machine learning. Using these definitions, we analyze the considered GNN-based recommender architectures to assess how and to what extent they encode such properties. Building on this analysis, we derive an explanatory framework that links measurable dataset characteristics to model behavior and performance. Taken together, this monograph re-frames GNN-based recommendation through its topological underpinnings and outlines open theoretical, data-centric, and evaluation challenges for the next generation of topology-aware recommender systems.

Paperid: 859, https://arxiv.org/pdf/2512.07245.pdf

Abstract:
Textual explanations make image classifier decisions transparent by describing the prediction rationale in natural language. Large vision-language models can generate captions but are designed for general visual understanding, not classifier-specific reasoning. Existing zero-shot explanation methods align global image features with language, producing descriptions of what is visible rather than what drives the prediction. We propose TEXTER, which overcomes this limitation by isolating decision-critical features before alignment. TEXTER identifies the neurons contributing to the prediction and emphasizes the features encoded in those neurons -- i.e., the decision-critical features. It then maps these emphasized features into the CLIP feature space to retrieve textual explanations that reflect the model's reasoning. A sparse autoencoder further improves interpretability, particularly for Transformer architectures. Extensive experiments show that TEXTER generates more faithful and interpretable explanations than existing methods. The code will be publicly released.

Paperid: 860, https://arxiv.org/pdf/2512.06973.pdf

Abstract:
Control Barrier Functions (CBFs) have emerged as a powerful tool for enforcing safety in optimization-based controllers, and their integration with Signal Temporal Logic (STL) has enabled the specification-driven synthesis of complex robotic behaviors. However, existing CBF-STL approaches typically rely on fixed hyperparameters and myopic, per-time step optimization, which can lead to overly conservative behavior, infeasibility near tight input limits, and difficulty satisfying long-horizon STL tasks. To address these limitations, we propose a feasibility-aware learning framework that embeds trainable, time-varying High Order Control Barrier Functions (HOCBFs) into a differentiable Quadratic Program (dQP). Our approach provides a systematic procedure for constructing time-varying HOCBF constraints for a broad fragment of STL and introduces a unified robustness measure that jointly captures STL satisfaction, QP feasibility, and control-bound compliance. Three neural networks-InitNet, RefNet, and an extended BarrierNet-collaborate to generate reference inputs and adapt constraint-related hyperparameters automatically over time and across initial conditions, reducing conservativeness while maximizing robustness. The resulting controller achieves STL satisfaction with strictly feasible dQPs and requires no manual tuning. Simulation results demonstrate that the proposed framework maintains high STL robustness under tight input bounds and significantly outperforms fixed-parameter and non-adaptive baselines in complex environments.

Paperid: 861, https://arxiv.org/pdf/2512.06515.pdf

Abstract:
Current language model safety paradigms often fall short in emotionally charged or high-stakes settings, where refusal-only approaches may alienate users and naive compliance can amplify risk. We propose ProSocialAlign, a test-time, parameter-efficient framework that steers generation toward safe, empathetic, and value-aligned responses without retraining the base model. We formalize five human-centered objectives and cast safety as lexicographic constrained generation: first, applying hard constraints to eliminate harmful continuations; then optimizing for prosocial quality within the safe set. Our method combines (i) directional regulation, a harm-mitigation mechanism that subtracts a learned "harm vector" in parameter space, and (ii) preference-aware autoregressive reward modeling trained jointly across attributes with gradient conflict resolution, enabling fine-grained, user-controllable decoding. Empirical evaluations across five safety benchmarks demonstrate state-of-the-art performance, reducing unsafe leakage and boosting alignment to human values, with strong gains across multiple evaluation metrics. ProSocialAlign offers a robust and modular foundation for generating context-sensitive, safe, and human-aligned responses at inference time.

Paperid: 862, https://arxiv.org/pdf/2512.06457.pdf

Abstract:
Attention-based large language models (LLMs) have transformed modern AI applications, but the quadratic cost of self-attention imposes significant compute and memory overhead. Dynamic sparsity (DS) attention mitigates this, yet its hardware efficiency is limited by the added prediction stage and the heavy memory traffic it entails. To address these limitations, this paper proposes BitStopper, a fine-grained algorithm-architecture co-design that operates without a sparsity predictor. First, a bit-serial enable stage fusion (BESF) mechanism is proposed to reuse and minimize the memory access by progressively terminating trivial tokens and merging the prediction stage into the execution stage. Second, a lightweight and adaptive token selection (LATS) strategy is developed to work in concert with the bit-level sparsity speculation. Third, a bit-level asynchronous processing (BAP) strategy is employed to improve compute utilization during the on-demand bit-grained memory fetching. Finally, an elaborate architecture is designed to translate the theoretical complexity reduction into practical performance improvement. Extensive evaluations demonstrate that, compared to state-of-the-art (SOTA) Transformer accelerators, BitStopper achieves 2.03x and 1.89x speedups over Sanger and SOFA, respectively, while delivering 2.4x and 2.1x improvements in energy efficiency.

Paperid: 863, https://arxiv.org/pdf/2512.05754.pdf

Abstract:
The scalability of high-fidelity video diffusion models (VDMs) is constrained by two key sources of redundancy: the quadratic complexity of global spatio-temporal attention and the computational overhead of long iterative denoising trajectories. Existing accelerators -- such as sparse attention and step-distilled samplers -- typically target a single dimension in isolation and quickly encounter diminishing returns, as the remaining bottlenecks become dominant. In this work, we introduce USV (Unified Sparsification for Video diffusion models), an end-to-end trainable framework that overcomes this limitation by jointly orchestrating sparsification across both the model's internal computation and its sampling process. USV learns a dynamic, data- and timestep-dependent sparsification policy that prunes redundant attention connections, adaptively merges semantically similar tokens, and reduces denoising steps, treating them not as independent tricks but as coordinated actions within a single optimization objective. This multi-dimensional co-design enables strong mutual reinforcement among previously disjoint acceleration strategies. Extensive experiments on large-scale video generation benchmarks demonstrate that USV achieves up to 83.3% speedup in the denoising process and 22.7% end-to-end acceleration, while maintaining high visual fidelity. Our results highlight unified, dynamic sparsification as a practical path toward efficient, high-quality video generation.

Paperid: 864, https://arxiv.org/pdf/2512.05745.pdf

Abstract:
Multimodal Large Language Models (MLLMs) are increasingly vulnerable to multimodal Indirect Prompt Injection (IPI) attacks, which embed malicious instructions in images, videos, or audio to hijack model behavior. Existing defenses, designed primarily for text-only LLMs, are unsuitable for countering these multimodal threats, as they are easily bypassed, modality-dependent, or generalize poorly. Inspired by activation steering researches, we hypothesize that a robust, general defense independent of modality can be achieved by steering the model's behavior in the representation space. Through extensive experiments, we discover that the instruction-following behavior of MLLMs is encoded in a subspace. Steering along directions within this subspace can enforce adherence to user instructions, forming the basis of a defense. However, we also found that a naive defense direction could be coupled with a utility-degrading direction, and excessive intervention strength harms model performance. To address this, we propose ARGUS, which searches for an optimal defense direction within the safety subspace that decouples from the utility degradation direction, further combining adaptive strength steering to achieve a better safety-utility trade-off. ARGUS also introduces lightweight injection detection stage to activate the defense on-demand, and a post-filtering stage to verify defense success. Experimental results show that ARGUS can achieve robust defense against multimodal IPI while maximally preserving the MLLM's utility.

Paperid: 865, https://arxiv.org/pdf/2512.04985.pdf

Abstract:
Guided or controlled data generation with diffusion models\blfootnote{Partial preliminary results of this work appeared in International Conference on Machine Learning 2025 \citep{li2025provable}.} has become a cornerstone of modern generative modeling. Despite substantial advances in diffusion model theory, the theoretical understanding of guided diffusion samplers remains severely limited. We make progress by developing a unified algorithmic and theoretical framework that accommodates both diffusion guidance and reward-guided diffusion. Aimed at fine-tuning diffusion models to improve certain rewards, we propose injecting a reward guidance term -- constructed from the difference between the original and reward-reweighted scores -- into the backward diffusion process, and rigorously quantify the resulting reward improvement over the unguided counterpart. As a key application, our framework shows that classifier-free guidance (CFG) decreases the expected reciprocal of the classifier probability, providing the first theoretical characterization of the specific performance metric that CFG improves for general target distributions. When applied to reward-guided diffusion, our framework yields a new sampler that is easy-to-train and requires no full diffusion trajectories during training. Numerical experiments further corroborate our theoretical findings.

Paperid: 866, https://arxiv.org/pdf/2512.04742.pdf

Abstract:
Rotatable antenna (RA) is a promising technology that can exploit new spatial degrees-of-freedom (DoFs) by flexibly adjusting the three-dimensional (3D) boresight direction of antennas. In this letter, we investigate an RA-enhanced cell-free system for downlink transmission, where multiple RA-equipped access points (APs) cooperatively serve multiple single-antenna users over the same time-frequency resource. Specifically, we aim to maximize the sum rate of all users by jointly optimizing the AP-user associations and the RA boresight directions. Accordingly, we propose a two-stage strategy to solve the AP-user association problem, and then employ fractional programming (FP) and successive convex approximation (SCA) techniques to optimize the RA boresight directions. Numerical results demonstrate that the proposed RA-enhanced cell-free system significantly outperforms various benchmark schemes.

Paperid: 867, https://arxiv.org/pdf/2512.04551.pdf

Abstract:
Speech emotion recognition (SER) is an important technology in human-computer interaction. However, achieving high performance is challenging due to emotional complexity and scarce annotated data. To tackle these challenges, we propose a multi-loss learning (MLL) framework integrating an energy-adaptive mixup (EAM) method and a frame-level attention module (FLAM). The EAM method leverages SNR-based augmentation to generate diverse speech samples capturing subtle emotional variations. FLAM enhances frame-level feature extraction for multi-frame emotional cues. Our MLL strategy combines Kullback-Leibler divergence, focal, center, and supervised contrastive loss to optimize learning, address class imbalance, and improve feature separability. We evaluate our method on four widely used SER datasets: IEMOCAP, MSP-IMPROV, RAVDESS, and SAVEE. The results demonstrate our method achieves state-of-the-art performance, suggesting its effectiveness and robustness.

Paperid: 868, https://arxiv.org/pdf/2512.04537.pdf

Abstract:
The advancement of embodied AI has unlocked significant potential for intelligent humanoid robots. However, progress in both Vision-Language-Action (VLA) models and world models is severely hampered by the scarcity of large-scale, diverse training data. A promising solution is to "robotize" web-scale human videos, which has been proven effective for policy training. However, these solutions mainly "overlay" robot arms to egocentric videos, which cannot handle complex full-body motions and scene occlusions in third-person videos, making them unsuitable for robotizing humans. To bridge this gap, we introduce X-Humanoid, a generative video editing approach that adapts the powerful Wan 2.2 model into a video-to-video structure and finetunes it for the human-to-humanoid translation task. This finetuning requires paired human-humanoid videos, so we designed a scalable data creation pipeline, turning community assets into 17+ hours of paired synthetic videos using Unreal Engine. We then apply our trained model to 60 hours of the Ego-Exo4D videos, generating and releasing a new large-scale dataset of over 3.6 million "robotized" humanoid video frames. Quantitative analysis and user studies confirm our method's superiority over existing baselines: 69% of users rated it best for motion consistency, and 62.1% for embodiment correctness.

Paperid: 869, https://arxiv.org/pdf/2512.04443.pdf

Abstract:
Spiking Neural Networks (SNNs) offer a promising and energy-efficient alternative to conventional neural networks, thanks to their sparse binary activation. However, they face challenges regarding memory and computation overhead due to complex spatio-temporal dynamics and the necessity for multiple backpropagation computations across timesteps during training. To mitigate this overhead, compression techniques such as quantization are applied to SNNs. Yet, naively applying quantization to SNNs introduces a mismatch in membrane potential, a crucial factor for the firing of spikes, resulting in accuracy degradation. In this paper, we introduce Membrane-aware Distillation on quantized Spiking Neural Network (MD-SNN), which leverages membrane potential to mitigate discrepancies after weight, membrane potential, and batch normalization quantization. To our knowledge, this study represents the first application of membrane potential knowledge distillation in SNNs. We validate our approach on various datasets, including CIFAR10, CIFAR100, N-Caltech101, and TinyImageNet, demonstrating its effectiveness for both static and dynamic data scenarios. Furthermore, for hardware efficiency, we evaluate the MD-SNN with SpikeSim platform, finding that MD-SNNs achieve 14.85X lower energy-delay-area product (EDAP), 2.64X higher TOPS/W, and 6.19X higher TOPS/mm2 compared to floating point SNNs at iso-accuracy on N-Caltech101 dataset.

Paperid: 870, https://arxiv.org/pdf/2512.04308.pdf

Abstract:
Recent advances in large multimodal models have enabled new opportunities in embodied AI, particularly in robotic manipulation. These models have shown strong potential in generalization and reasoning, but achieving reliable and responsible robotic behavior in real-world settings remains an open challenge. In high-stakes environments, robotic agents must go beyond basic task execution to perform risk-aware reasoning, moral decision-making, and physically grounded planning. We introduce ResponsibleRobotBench, a systematic benchmark designed to evaluate and accelerate progress in responsible robotic manipulation from simulation to real world. This benchmark consists of 23 multi-stage tasks spanning diverse risk types, including electrical, chemical, and human-related hazards, and varying levels of physical and planning complexity. These tasks require agents to detect and mitigate risks, reason about safety, plan sequences of actions, and engage human assistance when necessary. Our benchmark includes a general-purpose evaluation framework that supports multimodal model-based agents with various action representation modalities. The framework integrates visual perception, context learning, prompt construction, hazard detection, reasoning and planning, and physical execution. It also provides a rich multimodal dataset, supports reproducible experiments, and includes standardized metrics such as success rate, safety rate, and safe success rate. Through extensive experimental setups, ResponsibleRobotBench enables analysis across risk categories, task types, and agent configurations. By emphasizing physical reliability, generalization, and safety in decision-making, this benchmark provides a foundation for advancing the development of trustworthy, real-world responsible dexterous robotic systems. https://sites.google.com/view/responsible-robotbench

Paperid: 871, https://arxiv.org/pdf/2512.04085.pdf

Abstract:
We introduce the "single-life" learning paradigm, where we train a distinct vision model exclusively on egocentric videos captured by one individual. We leverage the multiple viewpoints naturally captured within a single life to learn a visual encoder in a self-supervised manner. Our experiments demonstrate three key findings. First, models trained independently on different lives develop a highly aligned geometric understanding. We demonstrate this by training visual encoders on distinct datasets each capturing a different life, both indoors and outdoors, as well as introducing a novel cross-attention-based metric to quantify the functional alignment of the internal representations developed by different models. Second, we show that single-life models learn generalizable geometric representations that effectively transfer to downstream tasks, such as depth estimation, in unseen environments. Third, we demonstrate that training on up to 30 hours from one week of the same person's life leads to comparable performance to training on 30 hours of diverse web data, highlighting the strength of single-life representation learning. Overall, our results establish that the shared structure of the world, both leads to consistency in models trained on individual lives, and provides a powerful signal for visual representation learning.

Paperid: 872, https://arxiv.org/pdf/2512.03874.pdf

Abstract:
Dexterous grasp generation aims to produce grasp poses that align with task requirements and human interpretable grasp semantics. However, achieving semantically controllable dexterous grasp synthesis remains highly challenging due to the lack of unified modeling of multiple semantic dimensions, including grasp taxonomy, contact semantics, and functional affordance. To address these limitations, we present OmniDexVLG, a multimodal, semantics aware grasp generation framework capable of producing structurally diverse and semantically coherent dexterous grasps under joint language and visual guidance. Our approach begins with OmniDexDataGen, a semantic rich dexterous grasp dataset generation pipeline that integrates grasp taxonomy guided configuration sampling, functional affordance contact point sampling, taxonomy aware differential force closure grasp sampling, and physics based optimization and validation, enabling systematic coverage of diverse grasp types. We further introduce OmniDexReasoner, a multimodal grasp type semantic reasoning module that leverages multi agent collaboration, retrieval augmented generation, and chain of thought reasoning to infer grasp related semantics and generate high quality annotations that align language instructions with task specific grasp intent. Building upon these components, we develop a unified Vision Language Grasping generation model that explicitly incorporates grasp taxonomy, contact structure, and functional affordance semantics, enabling fine grained control over grasp synthesis from natural language instructions. Extensive experiments in simulation and real world object grasping and ablation studies demonstrate that our method substantially outperforms state of the art approaches in terms of grasp diversity, contact semantic diversity, functional affordance diversity, and semantic consistency.

Paperid: 873, https://arxiv.org/pdf/2512.03394.pdf

Abstract:
Graph classification is a fundamental task in domains ranging from molecular property prediction to materials design. While graph neural networks (GNNs) achieve strong performance by learning expressive representations via message passing, they incur high computational costs, limiting their scalability and deployment on resource-constrained devices. Hyperdimensional Computing (HDC), also known as Vector Symbolic Architectures (VSA), offers a lightweight, brain-inspired alternative, yet existing HDC-based graph methods typically struggle to match the predictive performance of GNNs. In this work, we propose VS-Graph, a vector-symbolic graph learning framework that narrows the gap between the efficiency of HDC and the expressive power of message passing. VS-Graph introduces a Spike Diffusion mechanism for topology-driven node identification and an Associative Message Passing scheme for multi-hop neighborhood aggregation entirely within the high-dimensional vector space. Without gradient-based optimization or backpropagation, our method achieves competitive accuracy with modern GNNs, outperforming the prior HDC baseline by 4-5% on standard benchmarks such as MUTAG and DD. It also matches or exceeds the performance of the GNN baselines on several datasets while accelerating the training by a factor of up to 450x. Furthermore, VS-Graph maintains high accuracy even with the hypervector dimensionality reduced to D=128, demonstrating robustness under aggressive dimension compression and paving the way for ultra-efficient execution on edge and neuromorphic hardware.

Paperid: 874, https://arxiv.org/pdf/2512.02719.pdf

Abstract:
Large language models (LLMs) excel at explicit reasoning, but their implicit computational strategies remain underexplored. Decades of psychophysics research show that humans intuitively process and integrate noisy signals using near-optimal Bayesian strategies in perceptual tasks. We ask whether LLMs exhibit similar behaviour and perform optimal multimodal integration without explicit training or instruction. Adopting the psychophysics paradigm, we infer computational principles of LLMs from systematic behavioural studies. We introduce a behavioural benchmark - BayesBench: four magnitude estimation tasks (length, location, distance, and duration) over text and image, inspired by classic psychophysics, and evaluate a diverse set of nine LLMs alongside human judgments for calibration. Through controlled ablations of noise, context, and instruction prompts, we measure performance, behaviour and efficiency in multimodal cue-combination. Beyond accuracy and efficiency metrics, we introduce a Bayesian Consistency Score that detects Bayes-consistent behavioural shifts even when accuracy saturates. Our results show that while capable models often adapt in Bayes-consistent ways, accuracy does not guarantee robustness. Notably, GPT-5 Mini achieves perfect text accuracy but fails to integrate visual cues efficiently. This reveals a critical dissociation between capability and strategy, suggesting accuracy-centric benchmarks may over-index on performance while missing brittle uncertainty handling. These findings reveal emergent principled handling of uncertainty and highlight the correlation between accuracy and Bayesian tendencies. We release our psychophysics benchmark and consistency metric (https://bayes-bench.github.io) as evaluation tools and to inform future multimodal architecture designs.

Paperid: 875, https://arxiv.org/pdf/2512.02710.pdf

Abstract:
Automatic medical report generation can greatly reduce the workload of doctors, but it is often unreliable for real-world deployment. Current methods can write formally fluent sentences but may be factually flawed, introducing serious medical errors known as clinical hallucinations, which make them untrustworthy for diagnosis. To bridge this gap, we introduce HiMed-RL, a Hierarchical Medical Reward Learning Framework designed to explicitly prioritize clinical quality. HiMed-RL moves beyond simple text matching by deconstructing reward learning into three synergistic levels: it first ensures linguistic fluency at the token-level, then enforces factual grounding at the concept-level by aligning key medical terms with expert knowledge, and finally assesses high-level diagnostic consistency at the semantic-level using a specialized LLM verifier. This hierarchical reward is implemented via a Human-inspired Dynamic Reward Adjustment, a strategy which first teaches the model to learn basic facts before progressing to more complex diagnostic reasoning. Experimentally, HiMed-3B achieves state-of-the-art performance on both in-domain and out-of-domain benchmarks, particularly on the latter, with an improvement of 12.1% over the second-best baseline. Our work provides a robust paradigm for generating reports that not only improve fluency but clinical fine-grained quality.

Paperid: 876, https://arxiv.org/pdf/2512.02699.pdf

Abstract:
In this paper, we present Modality-Importance-Guided Reasoning (MIGR), a framework designed to improve the reliability of reasoning-based multimodal emotion understanding in multimodal large language models. Although existing methods have advanced emotion understanding, they often suffer from reasoning drift: models gradually rely on their own generated text instead of multimodal evidence, and their explanations are overly shaped by visually initiated reasoning paths. To address these issues, we introduce Modality Importance (MI), a simple yet effective mechanism for identifying the emotion-dominant modality. Using MI, MIGR reorganizes reasoning sequences so that explanations begin from the modality most critical to the target emotion, preventing early reasoning from being misled by less informative cues. Our two-stage framework-comprising modality-aligned supervised fine-tuning and modality-aware reward optimization-encourages models to generate emotionally grounded, causally relevant, and coherence-preserving explanations. Experimental results on the DFEW benchmark show that MIGR substantially improves reasoning reliability, decreasing instances of correct predictions accompanied by emotionally inconsistent explanations from 18.10% to 7.37%. These results confirm the benefit of initiating reasoning from the emotion-dominant modality.

Paperid: 877, https://arxiv.org/pdf/2512.02280.pdf

Abstract:
Artificial intelligence has advanced rapidly across perception, language, reasoning, and multimodal domains. Yet despite these achievements, modern AI systems remain fundamentally limited in their ability to self-monitor, self-correct, and regulate their behavior autonomously in dynamic contexts. This paper identifies and analyzes seven core deficiencies that constrain contemporary AI models: the absence of intrinsic self-monitoring, lack of meta-cognitive awareness, fixed and non-adaptive learning mechanisms, inability to restructure goals, lack of representational maintenance, insufficient embodied feedback, and the absence of intrinsic agency. Alongside identifying these limitations, we also outline a forward-looking perspective on how AI may evolve beyond them through architectures that mirror neurocognitive principles. We argue that these structural limitations prevent current architectures, including deep learning and transformer-based systems, from achieving robust generalization, lifelong adaptability, and real-world autonomy. Drawing on a comparative analysis of artificial systems and biological cognition [7], and integrating insights from AI research, cognitive science, and neuroscience, we outline how these capabilities are absent in current models and why scaling alone cannot resolve them. We conclude by advocating for a paradigmatic shift toward cognitively grounded AI (cognitive autonomy) capable of self-directed adaptation, dynamic representation management, and intentional, goal-oriented behavior, paired with reformative oversight mechanisms [8] that ensure autonomous systems remain interpretable, governable, and aligned with human values.

Paperid: 878, https://arxiv.org/pdf/2512.02027.pdf

Abstract:
Automatic transcription of stuttered speech remains a challenge, even for modern end-to-end (E2E) automatic speech recognition (ASR) frameworks. Dysfluencies and fluency-shaping artifacts are often overlooked, resulting in non-verbatim transcriptions with limited clinical and research value. We propose a parameter-efficient adaptation method to decode dysfluencies and fluency modifications as special tokens within transcriptions, evaluated on simulated (LibriStutter, English) and natural (KSoF, German) stuttered speech datasets. To mitigate ASR performance disparities and bias towards English, we introduce a multi-step fine-tuning strategy with language-adaptive pretraining. Tokenization analysis further highlights the tokenizer's English-centric bias, which poses challenges for improving performance on German data. Our findings demonstrate the effectiveness of lightweight adaptation techniques for dysfluency-aware ASR while exposing key limitations in multilingual E2E systems.

Paperid: 879, https://arxiv.org/pdf/2512.01996.pdf

Abstract:
Massively parallel simulation has reduced reinforcement learning (RL) training time for robots from days to minutes. However, achieving fast and reliable sim-to-real RL for humanoid control remains difficult due to the challenges introduced by factors such as high dimensionality and domain randomization. In this work, we introduce a simple and practical recipe based on off-policy RL algorithms, i.e., FastSAC and FastTD3, that enables rapid training of humanoid locomotion policies in just 15 minutes with a single RTX 4090 GPU. Our simple recipe stabilizes off-policy RL algorithms at massive scale with thousands of parallel environments through carefully tuned design choices and minimalist reward functions. We demonstrate rapid end-to-end learning of humanoid locomotion controllers on Unitree G1 and Booster T1 robots under strong domain randomization, e.g., randomized dynamics, rough terrain, and push perturbations, as well as fast training of whole-body human-motion tracking policies. We provide videos and open-source implementation at: https://younggyo.me/fastsac-humanoid.

Paperid: 880, https://arxiv.org/pdf/2512.01405.pdf

Abstract:
Foundation models (FMs) trained with different objectives and data learn diverse representations, making some more effective than others for specific downstream tasks. Existing adaptation strategies, such as parameter-efficient fine-tuning, focus on individual models and do not exploit the complementary strengths across models. Probing methods offer a promising alternative by extracting information from frozen models, but current techniques do not scale well with large feature sets and often rely on dataset-specific hyperparameter tuning. We propose Combined backBones (ComBo), a simple and scalable probing-based adapter that effectively integrates features from multiple models and layers. ComBo compresses activations from layers of one or more FMs into compact token-wise representations and processes them with a lightweight transformer for task-specific prediction. Crucially, ComBo does not require dataset-specific tuning or backpropagation through the backbone models. However, not all models are equally relevant for all tasks. To address this, we introduce a mechanism that leverages ComBo's joint multi-backbone probing to efficiently evaluate each backbone's task-relevance, enabling both practical model comparison and improved performance through selective adaptation. On the 19 tasks of the VTAB-1k benchmark, ComBo outperforms previous probing methods, matches or surpasses more expensive alternatives, such as distillation-based model merging, and enables efficient probing of tuned models. Our results demonstrate that ComBo offers a practical and general-purpose framework for combining diverse representations from multiple FMs.

Paperid: 881, https://arxiv.org/pdf/2512.01094.pdf

Abstract:
Masked Generative Models (MGM)s demonstrate strong capabilities in generating high-fidelity images. However, they need many sampling steps to create high-quality generations, resulting in slow inference speed. In this work, we propose Speed-RL, a novel paradigm for accelerating a pretrained MGMs to generate high-quality images in fewer steps. Unlike conventional distillation methods which formulate the acceleration problem as a distribution matching problem, where a few-step student model is trained to match the distribution generated by a many-step teacher model, we consider this problem as a reinforcement learning problem. Since the goal of acceleration is to generate high quality images in fewer steps, we can combine a quality reward with a speed reward and finetune the base model using reinforcement learning with the combined reward as the optimization target. Through extensive experiments, we show that the proposed method was able to accelerate the base model by a factor of 3x while maintaining comparable image quality.

Paperid: 882, https://arxiv.org/pdf/2512.00833.pdf

Abstract:
Modern circuits face various threats like reverse engineering, theft of intellectual property (IP), side-channel attacks, etc. Here, we present a novel approach for IP protection based on logic encryption (LE). Unlike established schemes for logic locking, our work obfuscates the circuit's structure and functionality by encoding and encrypting the logic itself. We devise an end-to-end method for practical LE implementation based on standard cryptographic algorithms, key-bit randomization, simple circuit design techniques, and system-level synthesis operations, all in a correct-by-construction manner. Our extensive analysis demonstrates the remarkable efficacy of our scheme, outperforming prior art against a range of oracle-less attacks covering crucial threat vectors, all with lower design overheads. We provide a full open-source release.

Paperid: 883, https://arxiv.org/pdf/2512.07538.pdf

Abstract:
Recognizing semantic differences across documents, especially in different languages, is crucial for text generation evaluation and multilingual content alignment. However, as a standalone task it has received little attention. We address this by introducing SwissGov-RSD, the first naturalistic, document-level, cross-lingual dataset for semantic difference recognition. It encompasses a total of 224 multi-parallel documents in English-German, English-French, and English-Italian with token-level difference annotations by human annotators. We evaluate a variety of open-source and closed source large language models as well as encoder models across different fine-tuning settings on this new benchmark. Our results show that current automatic approaches perform poorly compared to their performance on monolingual, sentence-level, and synthetic benchmarks, revealing a considerable gap for both LLMs and encoder models. We make our code and datasets publicly available.

Paperid: 884, https://arxiv.org/pdf/2512.07515.pdf

Abstract:
Detecting hallucinations in Retrieval-Augmented Generation (RAG) remains a challenge. Prior approaches attribute hallucinations to a binary conflict between internal knowledge (stored in FFNs) and retrieved context. However, this perspective is incomplete, failing to account for the impact of other components in the generative process, such as the user query, previously generated tokens, the current token itself, and the final LayerNorm adjustment. To address this, we introduce SPAD. First, we mathematically attribute each token's probability into seven distinct sources: Query, RAG, Past, Current Token, FFN, Final LayerNorm, and Initial Embedding. This attribution quantifies how each source contributes to the generation of the current token. Then, we aggregate these scores by POS tags to quantify how different components drive specific linguistic categories. By identifying anomalies, such as Nouns relying on Final LayerNorm, SPAD effectively detects hallucinations. Extensive experiments demonstrate that SPAD achieves state-of-the-art performance

Paperid: 885, https://arxiv.org/pdf/2512.06869.pdf

Abstract:
Large Language Models (LLMs) have achieved remarkable performance on single-turn tasks, yet their effectiveness deteriorates in multi-turn conversations. We define this phenomenon as cumulative contextual decay - a progressive degradation of contextual integrity caused by attention pollution, dilution, and drift. To address this challenge, we propose Rhea (Role-aware Heuristic Episodic Attention), a novel framework that decouples conversation history into two functionally independent memory modules: (1) an Instructional Memory (IM) that persistently stores high-fidelity global constraints via a structural priority mechanism, and (2) an Episodic Memory (EM) that dynamically manages user-model interactions via asymmetric noise control and heuristic context retrieval. During inference, Rhea constructs a high signal-to-noise context by applying its priority attention: selectively integrating relevant episodic information while always prioritizing global instructions. To validate this approach, experiments on multiple multi-turn conversation benchmarks - including MT-Eval and Long-MT-Bench+ - show that Rhea mitigates performance decay and improves overall accuracy by 1.04 points on a 10-point scale (a 16% relative gain over strong baselines). Moreover, Rhea maintains near-perfect instruction fidelity (IAR > 8.1) across long-horizon interactions. These results demonstrate that Rhea provides a principled and effective framework for building more precise, instruction-consistent conversational LLMs.

Paperid: 886, https://arxiv.org/pdf/2512.06571.pdf

Abstract:
Learning fast and robust ball-kicking skills is a critical capability for humanoid soccer robots, yet it remains a challenging problem due to the need for rapid leg swings, postural stability on a single support foot, and robustness under noisy sensory input and external perturbations (e.g., opponents). This paper presents a reinforcement learning (RL)-based system that enables humanoid robots to execute robust continual ball-kicking with adaptability to different ball-goal configurations. The system extends a typical teacher-student training framework -- in which a "teacher" policy is trained with ground truth state information and the "student" learns to mimic it with noisy, imperfect sensing -- by including four training stages: (1) long-distance ball chasing (teacher); (2) directional kicking (teacher); (3) teacher policy distillation (student); and (4) student adaptation and refinement (student). Key design elements -- including tailored reward functions, realistic noise modeling, and online constrained RL for adaptation and refinement -- are critical for closing the sim-to-real gap and sustaining performance under perceptual uncertainty. Extensive evaluations in both simulation and on a real robot demonstrate strong kicking accuracy and goal-scoring success across diverse ball-goal configurations. Ablation studies further highlight the necessity of the constrained RL, noise modeling, and the adaptation stage. This work presents a system for learning robust continual humanoid ball-kicking under imperfect perception, establishing a benchmark task for visuomotor skill learning in humanoid whole-body control.

Paperid: 887, https://arxiv.org/pdf/2512.06553.pdf

Abstract:
We propose a statistical framework built on latent variable modeling for scaling laws of large language models (LLMs). Our work is motivated by the rapid emergence of numerous new LLM families with distinct architectures and training strategies, evaluated on an increasing number of benchmarks. This heterogeneity makes a single global scaling curve inadequate for capturing how performance varies across families and benchmarks. To address this, we propose a latent variable modeling framework in which each LLM family is associated with a latent variable that captures the common underlying features in that family. An LLM's performance on different benchmarks is then driven by its latent skills, which are jointly determined by the latent variable and the model's own observable features. We develop an estimation procedure for this latent variable model and establish its statistical properties. We also design efficient numerical algorithms that support estimation and various downstream tasks. Empirically, we evaluate the approach on 12 widely used benchmarks from the Open LLM Leaderboard (v1/v2).

Paperid: 888, https://arxiv.org/pdf/2512.06201.pdf

Abstract:
We introduce K2-V2, a 360-open LLM built from scratch as a superior base for reasoning adaptation, in addition to functions such as conversation and knowledge retrieval from general LLMs. It stands as the strongest fully open model, rivals open-weight leaders in its size class, outperforms Qwen2.5-72B and approaches the performance of Qwen3-235B. We actively infuse domain knowledge, reasoning, long-context, and tool use throughout the training process. This explicitly prepares the model for complex reasoning tasks. We demonstrate this potential using simple supervised fine-tuning, establishing a strong baseline that indicates significant headroom for advanced alignment. By releasing the full training history and data composition, we maximize the effectiveness of continuous training, a key open source production scenario. We release the model weights and signature LLM360 artifacts, such as complete training data, to empower the community with a capable, reasoning-centric foundation.

Paperid: 889, https://arxiv.org/pdf/2512.05343.pdf

Abstract:
Generative methods for 3D assets have recently achieved remarkable progress, yet providing intuitive and precise control over the object geometry remains a key challenge. Existing approaches predominantly rely on text or image prompts, which often fall short in geometric specificity: language can be ambiguous, and images are cumbersome to edit. In this work, we introduce SpaceControl, a training-free test-time method for explicit spatial control of 3D generation. Our approach accepts a wide range of geometric inputs, from coarse primitives to detailed meshes, and integrates seamlessly with modern pre-trained generative models without requiring any additional training. A controllable parameter lets users trade off between geometric fidelity and output realism. Extensive quantitative evaluation and user studies demonstrate that SpaceControl outperforms both training-based and optimization-based baselines in geometric faithfulness while preserving high visual quality. Finally, we present an interactive user interface that enables online editing of superquadrics for direct conversion into textured 3D assets, facilitating practical deployment in creative workflows. Find our project page at https://spacecontrol3d.github.io/

Paperid: 890, https://arxiv.org/pdf/2512.04677.pdf

Abstract:
Existing diffusion-based video generation methods are fundamentally constrained by sequential computation and long-horizon inconsistency, limiting their practical adoption in real-time, streaming audio-driven avatar synthesis. We present Live Avatar, an algorithm-system co-designed framework that enables efficient, high-fidelity, and infinite-length avatar generation using a 14-billion-parameter diffusion model. Our approach introduces Timestep-forcing Pipeline Parallelism (TPP), a distributed inference paradigm that pipelines denoising steps across multiple GPUs, effectively breaking the autoregressive bottleneck and ensuring stable, low-latency real-time streaming. To further enhance temporal consistency and mitigate identity drift and color artifacts, we propose the Rolling Sink Frame Mechanism (RSFM), which maintains sequence fidelity by dynamically recalibrating appearance using a cached reference image. Additionally, we leverage Self-Forcing Distribution Matching Distillation to facilitate causal, streamable adaptation of large-scale models without sacrificing visual quality. Live Avatar demonstrates state-of-the-art performance, reaching 20 FPS end-to-end generation on 5 H800 GPUs, and, to the best of our knowledge, is the first to achieve practical, real-time, high-fidelity avatar generation at this scale. Our work establishes a new paradigm for deploying advanced diffusion models in industrial long-form video synthesis applications.

Paperid: 891, https://arxiv.org/pdf/2512.04601.pdf

Abstract:
Large language model (LLM) agents -- LLMs that dynamically interact with an environment over long horizons -- have become an increasingly important area of research, enabling automation in complex tasks involving tool-use, web browsing, and dialogue with people. In the absence of expert demonstrations, training LLM agents has relied on policy gradient methods that optimize LLM policies with respect to an (often sparse) reward function. However, in long-horizon tasks with sparse rewards, learning from trajectory-level rewards can be noisy, leading to training that is unstable and has high sample complexity. Furthermore, policy improvement hinges on discovering better actions through exploration, which can be difficult when actions lie in natural language space. In this paper, we propose Natural Language Actor-Critic (NLAC), a novel actor-critic algorithm that trains LLM policies using a generative LLM critic that produces natural language rather than scalar values. This approach leverages the inherent strengths of LLMs to provide a richer and more actionable training signal; particularly, in tasks with large, open-ended action spaces, natural language explanations for why an action is suboptimal can be immensely useful for LLM policies to reason how to improve their actions, without relying on random exploration. Furthermore, our approach can be trained off-policy without policy gradients, offering a more data-efficient and stable alternative to existing on-policy methods. We present results on a mixture of reasoning, web browsing, and tool-use with dialogue tasks, demonstrating that NLAC shows promise in outperforming existing training approaches and offers a more scalable and stable training paradigm for LLM agents.

Paperid: 892, https://arxiv.org/pdf/2512.04564.pdf

Abstract:
Supervised deep learning (DL) receives great interest for automated analysis of microscopic images with an increasing body of literature supporting its potential. The development and validation of those DL models relies heavily on the availability of high-quality, large-scale datasets. However, creating such datasets is a complex and resource-intensive process, often hindered by challenges such as time constraints, domain variability, and risks of bias in image collection and label creation. This review provides a comprehensive guide to the critical steps in dataset creation, including: 1) image acquisition, 2) selection of annotation software, and 3) annotation creation. In addition to ensuring a sufficiently large number of images, it is crucial to address sources of image variability (domain shifts) - such as those related to slide preparation and digitization - that could lead to algorithmic errors if not adequately represented in the training data. Key quality criteria for annotations are the three "C"s: correctness, completeness, and consistency. This review explores methods to enhance annotation quality through the use of advanced techniques that mitigate the limitations of single annotators. To support dataset creators, a standard operating procedure (SOP) is provided as supplemental material, outlining best practices for dataset development. Furthermore, the article underscores the importance of open datasets in driving innovation and enhancing reproducibility of DL research. By addressing the challenges and offering practical recommendations, this review aims to advance the creation of and availability to high-quality, large-scale datasets, ultimately contributing to the development of generalizable and robust DL models for pathology applications.

Paperid: 893, https://arxiv.org/pdf/2512.04563.pdf

Abstract:
Visual Spatial Reasoning is crucial for enabling Multimodal Large Language Models (MLLMs) to understand object properties and spatial relationships, yet current models still struggle with 3D-aware reasoning. Existing approaches typically enhance either perception, by augmenting RGB inputs with auxiliary modalities such as depth and segmentation, or reasoning, by training on spatial VQA datasets and applying reinforcement learning, and thus treat these two aspects in isolation. In this work, we investigate whether a unified MLLM can develop an intrinsic ability to enhance spatial perception and, through adaptive interleaved reasoning, achieve stronger spatial intelligence. We propose \textbf{COOPER}, a unified MLLM that leverages depth and segmentation as auxiliary modalities and is trained in two stages to acquire auxiliary modality generation and adaptive, interleaved reasoning capabilities. COOPER achieves an average \textbf{6.91\%} improvement in spatial reasoning while maintaining general performance. Moreover, even a variant trained only for auxiliary modality generation attains a \textbf{7.92\%} gain on distance and size estimation, suggesting that learning to generate auxiliary modalities helps internalize spatial knowledge and strengthen spatial understanding.

Paperid: 894, https://arxiv.org/pdf/2512.04519.pdf

Abstract:
Autoregressive (AR) diffusion enables streaming, interactive long-video generation by producing frames causally, yet maintaining coherence over minute-scale horizons remains challenging due to accumulated errors, motion drift, and content repetition. We approach this problem from a memory perspective, treating video synthesis as a recurrent dynamical process that requires coordinated short- and long-term context. We propose VideoSSM, a Long Video Model that unifies AR diffusion with a hybrid state-space memory. The state-space model (SSM) serves as an evolving global memory of scene dynamics across the entire sequence, while a context window provides local memory for motion cues and fine details. This hybrid design preserves global consistency without frozen, repetitive patterns, supports prompt-adaptive interaction, and scales in linear time with sequence length. Experiments on short- and long-range benchmarks demonstrate state-of-the-art temporal consistency and motion stability among autoregressive video generator especially at minute-scale horizons, enabling content diversity and interactive prompt-based control, thereby establishing a scalable, memory-aware framework for long video generation.

Paperid: 895, https://arxiv.org/pdf/2512.04238.pdf

Abstract:
Vision-language models are increasingly integrated into clinical workflows. However, existing benchmarks primarily assess performance on common anatomical presentations and fail to capture the challenges posed by rare variants. To address this gap, we introduce AdversarialAnatomyBench, the first benchmark comprising naturally occurring rare anatomical variants across diverse imaging modalities and anatomical regions. We call such variants that violate learned priors about "typical" human anatomy natural adversarial anatomy. Benchmarking 22 state-of-the-art VLMs with AdversarialAnatomyBench yielded three key insights. First, when queried with basic medical perception tasks, mean accuracy dropped from 74% on typical to 29% on atypical anatomy. Even the best-performing models, GPT-5, Gemini 2.5 Pro, and Llama 4 Maverick, showed performance drops of 41-51%. Second, model errors closely mirrored expected anatomical biases. Third, neither model scaling nor interventions, including bias-aware prompting and test-time reasoning, resolved these issues. These findings highlight a critical and previously unquantified limitation in current VLM: their poor generalization to rare anatomical presentations. AdversarialAnatomyBench provides a foundation for systematically measuring and mitigating anatomical bias in multimodal medical AI systems.

Paperid: 896, https://arxiv.org/pdf/2512.04110.pdf

Abstract:
In Fall 2023, we introduced a new AI Literacy class called The Essentials of AI for Life and Society (CS 109), a one-credit, seminar course consisting mainly of guest lectures, which was open to the entire university, including students, staff, and faculty. Building on its success and popularity, this paper describes our significant expansion of the course into a full-scale three-credit undergraduate course (CS 309), with an expanded emphasis on student engagement, interactivity, and ethics-related components. To knit together content from the guest lecturers, we implemented a flipped classroom. This model used weekly asynchronous learning modules--integrating pre-recorded expert lectures, collaborative readings, and ethical reflections--which were then unified by the course instructor during a live, interactive discussion session. To maintain the broad accessibility of the material (no prerequisites), the course introduced substantive, non-programming homework assignments in which students applied AI concepts to grounded, real-world problems. This work culminated in a final project analyzing the ethical and societal implications of a chosen AI tool. The redesigned course received overwhelmingly positive student feedback, highlighting its interactivity, coherence, and accessible and engaging assignments. This paper details the course's evolution, its pedagogical structure, and the lessons learned in developing a core AI literacy course. All course materials are freely available for others to use and build upon.

Paperid: 897, https://arxiv.org/pdf/2512.04077.pdf

Abstract:
For a remote estimation system, we study age of incorrect information (AoII), which is a recently proposed semantic-aware freshness metric. In particular, we assume an information source observing a discrete-time finite-state Markov chain (DTMC) and employing push-based transmissions of status update packets towards the monitor which is tasked with remote estimation of the source. The source-to-monitor channel delay is assumed to have a general discrete-time phase-type (DPH) distribution, whereas the zero-delay reverse channel ensures that the source has perfect information on AoII and the remote estimate. A multi-threshold transmission policy is employed where packet transmissions are initiated when the AoII process exceeds a threshold which may be different for each estimation value. In this general setting, our goal is to minimize the weighted sum of time average of an arbitrary function of AoII and estimation, and transmission costs, by suitable choice of the thresholds. We formulate the problem as a semi-Markov decision process (SMDP) with the same state-space as the original DTMC to obtain the optimum multi-threshold policy whereas the parameters of the SMDP are obtained by using a novel stochastic tool called dual-regime absorbing Markov chain (DR-AMC), and its corresponding absorption time distribution named as dual-regime DPH (DR-DPH).

Paperid: 898, https://arxiv.org/pdf/2512.03783.pdf

Abstract:
Recent advances in Omni models have enabled unified multimodal perception and generation. However, most existing systems still exhibit rigid reasoning behaviors, either overthinking simple problems or failing to reason when necessary. To address this limitation, we propose Omni-AutoThink, a novel adaptive reasoning framework that dynamically adjusts the model's reasoning depth according to task difficulty. Our framework comprises two stages: (1) an Adaptive Supervised Fine-Tuning (Adaptive SFT) stage, which endows the Omni model with fundamental reasoning capability using large-scale reasoning-augmented data, and (2) an Adaptive Reinforcement Learning (Adaptive GRPO) stage, which optimizes reasoning behaviors based on task complexity and reward feedback. We further construct a comprehensive adaptive reasoning benchmark that spans text-only, text-audio, text-visual, and text-audio-visual modalities, providing both training and evaluation splits for multimodal reasoning assessment. Experimental results demonstrate that our proposed framework significantly improves adaptive reasoning performance compared to previous baselines. All benchmark data and code will be publicly released.

Paperid: 899, https://arxiv.org/pdf/2512.02580.pdf

Abstract:
Reinforcement learning has emerged as a paradigm for post-training large language models, boosting their reasoning capabilities. Such approaches compute an advantage value for each sample, reflecting better or worse performance than expected, thereby yielding both positive and negative signals for training. However, the indiscriminate mixing of the two signals in existing methods, especially from the early stages, may lead to ambiguous guidance and limited gains. To address this issue, we propose **CAPO** (**C**urriculum **A**dvantage **P**olicy **O**ptimization), an adaptive curriculum mechanism based on advantage signals. The proposed mechanism bootstraps imitation learning with positive-only advantage samples to establish robust foundations, and subsequently introduces negative signals to cultivate discriminative capabilities, thereby improving generalization across complex scenarios. Compatible with diverse optimization methods including GRPO, PPO, RLOO, and Reinforce++, our method consistently achieves stable and significant improvements in mathematical reasoning tasks, and further generalizes effectively to multimodal Graphical User Interface (GUI) reasoning scenarios, establishing itself as a versatile and robust optimization framework.

Paperid: 900, https://arxiv.org/pdf/2512.02551.pdf

Abstract:
In this paper, we propose CUDA-L2, a system that combines large language models (LLMs) and reinforcement learning (RL) to automatically optimize Half-precision General Matrix Multiply (HGEMM) CUDA kernels. Using CUDA execution speed as the RL reward, CUDA-L2 automatically optimizes HGEMM kernels across 1,000 configurations. CUDA-L2 systematically outperforms major matmul baselines to date, from the widely-used {\it torch.matmul} to state-of-the-art Nvidia's closed-source libraries, i.e., {\it cuBLAS}, {\it cuBLASLt}. In offline mode, where kernels are executed consecutively without time intervals, CUDA-L2 yields +22.0\% over {\it torch.matmul} on average; +19.2\% over {\it cuBLAS} using the optimal layout configuration (normal-normal NN and transposed-normal TN); +16.8\% over {\it cuBLASLt-heuristic}, which queries {\it cuBLASLt} library and selects the algorithm based on the heuristic's suggestion; and +11.4\% over the most competitive {\it cuBLASLt-AutoTuning} model, which selects the fastest algorithm from up to 100 candidates from {\it cuBLASLt}'s suggestions. In server mode, where kernels are executed at random intervals simulating real-time inference, the speedups further increase to +28.7\%, +26.0\%, +22.4\%, and +15.9\% for {\it torch.matmul}, {\it cuBLAS}, {\it cuBLASLt-heuristic}, and {\it cuBLASLt-AutoTuning} respectively. CUDA-L2 shows that even the most performance-critical, heavily-optimized kernels like HGEMM can be improved through LLM-guided RL automation by systematically exploring configuration spaces at scales impractical for humans. Project and code can be found at github.com/deepreinforce-ai/CUDA-L2

Paperid: 901, https://arxiv.org/pdf/2512.01773.pdf

Abstract:
The rise of generalist robotic policies has created an exponential demand for large-scale training data. However, on-robot data collection is labor-intensive and often limited to specific environments. In contrast, open-world images capture a vast diversity of real-world scenes that naturally align with robotic manipulation tasks, offering a promising avenue for low-cost, large-scale robot data acquisition. Despite this potential, the lack of associated robot actions hinders the practical use of open-world images for robot learning, leaving this rich visual resource largely unexploited. To bridge this gap, we propose IGen, a framework that scalably generates realistic visual observations and executable actions from open-world images. IGen first converts unstructured 2D pixels into structured 3D scene representations suitable for scene understanding and manipulation. It then leverages the reasoning capabilities of vision-language models to transform scene-specific task instructions into high-level plans and generate low-level actions as SE(3) end-effector pose sequences. From these poses, it synthesizes dynamic scene evolution and renders temporally coherent visual observations. Experiments validate the high quality of visuomotor data generated by IGen, and show that policies trained solely on IGen-synthesized data achieve performance comparable to those trained on real-world data. This highlights the potential of IGen to support scalable data generation from open-world images for generalist robotic policy training.

Paperid: 902, https://arxiv.org/pdf/2512.01675.pdf

Abstract:
Recent advances in text-to-image diffusion models enable high-fidelity generation across diverse prompts. However, these models falter in long-tail settings, such as medical imaging, where rare pathologies comprise a small fraction of data. This results in mode collapse: tail-class outputs lack quality and diversity, undermining the goal of synthetic data augmentation for underrepresented conditions. We pinpoint gradient conflicts between frequent head and rare tail classes as the primary culprit, a factor unaddressed by existing sampling or conditioning methods that mainly steer inference without altering the learned distribution. To resolve this, we propose GRASP: Guided Residual Adapters with Sample-wise Partitioning. GRASP uses external priors to statically partition samples into clusters that minimize intra-group gradient clashes. It then fine-tunes pre-trained models by injecting cluster-specific residual adapters into transformer feedforward layers, bypassing learned gating for stability and efficiency. On the long-tail MIMIC-CXR-LT dataset, GRASP yields superior FID and diversity metrics, especially for rare classes, outperforming baselines like vanilla fine-tuning and Mixture of Experts variants. Downstream classification on NIH-CXR-LT improves considerably for tail labels. Generalization to ImageNet-LT confirms broad applicability. Our method is lightweight, scalable, and readily integrates with diffusion pipelines.

Paperid: 903, https://arxiv.org/pdf/2512.00771.pdf

Abstract:
Robust 3D geometry estimation from videos is critical for applications such as autonomous navigation, SLAM, and 3D scene reconstruction. Recent methods like DUSt3R demonstrate that regressing dense pointmaps from image pairs enables accurate and efficient pose-free reconstruction. However, existing RGB-only approaches struggle under real-world conditions involving dynamic objects and extreme illumination, due to the inherent limitations of conventional cameras. In this paper, we propose EAG3R, a novel geometry estimation framework that augments pointmap-based reconstruction with asynchronous event streams. Built upon the MonST3R backbone, EAG3R introduces two key innovations: (1) a retinex-inspired image enhancement module and a lightweight event adapter with SNR-aware fusion mechanism that adaptively combines RGB and event features based on local reliability; and (2) a novel event-based photometric consistency loss that reinforces spatiotemporal coherence during global optimization. Our method enables robust geometry estimation in challenging dynamic low-light scenes without requiring retraining on night-time data. Extensive experiments demonstrate that EAG3R significantly outperforms state-of-the-art RGB-only baselines across monocular depth estimation, camera pose tracking, and dynamic reconstruction tasks.

Paperid: 904, https://arxiv.org/pdf/2512.00758.pdf

Abstract:
Movable antenna (MA) technology exhibits great promise for enhancing the sensing capabilities of future sixth-generation (6G) networks due to its capability to alter antenna array geometry. With the growing prevalence of near-field propagation at ultra-high frequencies, this paper focuses on the application of one-dimensional (1D) and two-dimensional (2D) MA arrays for near-field sensing to jointly estimate the angle and distance information about a target. First, for the 1D MA array scenario, to gain insights into MA-enhanced near-field sensing, we investigate two simplified cases with only angle-of-arrival (AoA) or distance estimation, respectively, assuming that the other information is already known. The worst-case Cramer-Rao bounds (CRBs) on the mean square errors (MSEs) of the AoA estimation and the distance estimation are derived in these two cases. Then, we jointly optimize the positions of the MAs within the 1D array to minimize these CRBs and derive their closed-form solutions, which yield an identical array geometry to MA-enhanced far-field sensing. For the more challenging joint AoA and distance estimation, since the associated worst-case CRB is a highly complex and non-convex function with respect to the MA positions, a discrete sampling-based approach is proposed to sequentially update the MA positions and obtain an efficient suboptimal solution. Furthermore, we investigate the worst-case CRB minimization problems for a 2D MA array under various conditions and extend our proposed algorithms to solve them efficiently. Numerical results demonstrate that the proposed MA-enhanced near-field sensing scheme dramatically outperforms conventional fixed-position antennas (FPAs). Moreover, the joint angle and distance estimation results in a different array geometry from that in the individual estimation of angle/distance or far-field sensing.

Paperid: 905, https://arxiv.org/pdf/2512.00020.pdf

Abstract:
Code generation has emerged as a critical research area at the intersection of Software Engineering (SE) and Artificial Intelligence (AI), attracting significant attention from both academia and industry. Within this broader landscape, Verilog, as a representative hardware description language (HDL), plays a fundamental role in digital circuit design and verification, making its automated generation particularly significant for Electronic Design Automation (EDA). Consequently, recent research has increasingly focused on applying Large Language Models (LLMs) to Verilog code generation, particularly at the Register Transfer Level (RTL), exploring how these AI-driven techniques can be effectively integrated into hardware design workflows. Despite substantial research efforts have explored LLM applications in this domain, a comprehensive survey synthesizing these developments remains absent from the literature. This review fill addresses this gap by providing a systematic literature review of LLM-based methods for Verilog code generation, examining their effectiveness, limitations, and potential for advancing automated hardware design. The review encompasses research work from conferences and journals in the fields of SE, AI, and EDA, encompassing 70 papers published on venues, along with 32 high-quality preprint papers, bringing the total to 102 papers. By answering four key research questions, we aim to (1) identify the LLMs used for Verilog generation, (2) examine the datasets and metrics employed in evaluation, (3) categorize the techniques proposed for Verilog generation, and (4) analyze LLM alignment approaches for Verilog generation. Based on our findings, we have identified a series of limitations of existing studies. Finally, we have outlined a roadmap highlighting potential opportunities for future research endeavors in LLM-assisted hardware design.

Paperid: 906, https://arxiv.org/pdf/2512.07829.pdf

Abstract:
Visual generative models (e.g., diffusion models) typically operate in compressed latent spaces to balance training efficiency and sample quality. In parallel, there has been growing interest in leveraging high-quality pre-trained visual representations, either by aligning them inside VAEs or directly within the generative model. However, adapting such representations remains challenging due to fundamental mismatches between understanding-oriented features and generation-friendly latent spaces. Representation encoders benefit from high-dimensional latents that capture diverse hypotheses for masked regions, whereas generative models favor low-dimensional latents that must faithfully preserve injected noise. This discrepancy has led prior work to rely on complex objectives and architectures. In this work, we propose FAE (Feature Auto-Encoder), a simple yet effective framework that adapts pre-trained visual representations into low-dimensional latents suitable for generation using as little as a single attention layer, while retaining sufficient information for both reconstruction and understanding. The key is to couple two separate deep decoders: one trained to reconstruct the original feature space, and a second that takes the reconstructed features as input for image generation. FAE is generic; it can be instantiated with a variety of self-supervised encoders (e.g., DINO, SigLIP) and plugged into two distinct generative families: diffusion models and normalizing flows. Across class-conditional and text-to-image benchmarks, FAE achieves strong performance. For example, on ImageNet 256x256, our diffusion model with CFG attains a near state-of-the-art FID of 1.29 (800 epochs) and 1.70 (80 epochs). Without CFG, FAE reaches the state-of-the-art FID of 1.48 (800 epochs) and 2.08 (80 epochs), demonstrating both high quality and fast learning.

Paperid: 907, https://arxiv.org/pdf/2512.07584.pdf

Abstract:
We introduce LongCat-Image, a pioneering open-source and bilingual (Chinese-English) foundation model for image generation, designed to address core challenges in multilingual text rendering, photorealism, deployment efficiency, and developer accessibility prevalent in current leading models. 1) We achieve this through rigorous data curation strategies across the pre-training, mid-training, and SFT stages, complemented by the coordinated use of curated reward models during the RL phase. This strategy establishes the model as a new state-of-the-art (SOTA), delivering superior text-rendering capabilities and remarkable photorealism, and significantly enhancing aesthetic quality. 2) Notably, it sets a new industry standard for Chinese character rendering. By supporting even complex and rare characters, it outperforms both major open-source and commercial solutions in coverage, while also achieving superior accuracy. 3) The model achieves remarkable efficiency through its compact design. With a core diffusion model of only 6B parameters, it is significantly smaller than the nearly 20B or larger Mixture-of-Experts (MoE) architectures common in the field. This ensures minimal VRAM usage and rapid inference, significantly reducing deployment costs. Beyond generation, LongCat-Image also excels in image editing, achieving SOTA results on standard benchmarks with superior editing consistency compared to other open-source works. 4) To fully empower the community, we have established the most comprehensive open-source ecosystem to date. We are releasing not only multiple model versions for text-to-image and image editing, including checkpoints after mid-training and post-training stages, but also the entire toolchain of training procedure. We believe that the openness of LongCat-Image will provide robust support for developers and researchers, pushing the frontiers of visual content creation.

Paperid: 908, https://arxiv.org/pdf/2512.07417.pdf

Abstract:
Effective traffic control is essential for mitigating congestion in transportation networks. Conventional traffic management strategies, including route guidance, ramp metering, and traffic signal control, often rely on state feedback controllers, used for their simplicity and reactivity; however, they lack the adaptability required to cope with complex and time-varying traffic dynamics. This paper proposes a multi-agent reinforcement learning framework in which each agent adaptively tunes the parameters of a state feedback traffic controller, combining the reactivity of state feedback controllers with the adaptability of reinforcement learning. By tuning parameters at a lower frequency rather than directly determining control actions at a high frequency, the reinforcement learning agents achieve improved training efficiency while maintaining adaptability to varying traffic conditions. The multi-agent structure further enhances system robustness, as local controllers can operate independently in the event of partial failures. The proposed framework is evaluated on a simulated multi-class transportation network under varying traffic conditions. Results show that the proposed multi-agent framework outperforms the no control and fixed-parameter state feedback control cases, while performing on par with the single-agent RL-based adaptive state feedback control, with a much better resilience to partial failures.

Paperid: 909, https://arxiv.org/pdf/2512.06956.pdf

Abstract:
Inverse reinforcement learning aims to infer the reward function that explains expert behavior observed through trajectories of state--action pairs. A long-standing difficulty in classical IRL is the non-uniqueness of the recovered reward: many reward functions can induce the same optimal policy, rendering the inverse problem ill-posed. In this paper, we develop a statistical framework for Inverse Entropy-regularized Reinforcement Learning that resolves this ambiguity by combining entropy regularization with a least-squares reconstruction of the reward from the soft Bellman residual. This combination yields a unique and well-defined so-called least-squares reward consistent with the expert policy. We model the expert demonstrations as a Markov chain with the invariant distribution defined by an unknown expert policy $π^\star$ and estimate the policy by a penalized maximum-likelihood procedure over a class of conditional distributions on the action space. We establish high-probability bounds for the excess Kullback--Leibler divergence between the estimated policy and the expert policy, accounting for statistical complexity through covering numbers of the policy class. These results lead to non-asymptotic minimax optimal convergence rates for the least-squares reward function, revealing the interplay between smoothing (entropy regularization), model complexity, and sample size. Our analysis bridges the gap between behavior cloning, inverse reinforcement learning, and modern statistical learning theory.

Paperid: 910, https://arxiv.org/pdf/2512.06784.pdf

Abstract:
The sparse activation mechanism of mixture of experts (MoE) model empowers edge intelligence with enhanced training efficiency and reduced computational resource consumption. However, traditional token routing in distributed MoE training faces significant challenges in resource-constrained edge networks characterized by heterogeneous computing capabilities and stochastic token arrivals, which inevitably suffer from workload backlog, resource inefficiency, and performance degradation. To address this issue, we propose a novel Lyapunov-based token routing framework for distributed MoE training over resource-heterogeneous edge networks, termed Stable-MoE. Specifically, we formulate a stochastic optimization problem to maximize both system throughput and gating consistency via optimizing the token routing strategy and computational resource allocation, while ensuring long-term stability of both token and energy queues at the edge devices. Using the Lyapunov optimization, we transform the intractable long-term optimization problem into tractable per-slot subproblems by enabling online decision-making of token routing and computation frequency utilization without the knowledge of future system states. Experimental results on the SVHN and CIFAR-100 datasets demonstrate that Stable-MoE outperforms the baselines with at least 40% and 5% gains in system throughput and test accuracy, respectively.

Paperid: 911, https://arxiv.org/pdf/2512.06204.pdf

Abstract:
How much does a trained RL policy actually use its past observations? We propose \emph{Temporal Range}, a model-agnostic metric that treats first-order sensitivities of multiple vector outputs across a temporal window to the input sequence as a temporal influence profile and summarizes it by the magnitude-weighted average lag. Temporal Range is computed via reverse-mode automatic differentiation from the Jacobian blocks $\partial y_s/\partial x_t\in\mathbb{R}^{c\times d}$ averaged over final timesteps $s\in\{t+1,\dots,T\}$ and is well-characterized in the linear setting by a small set of natural axioms. Across diagnostic and control tasks (POPGym; flicker/occlusion; Copy-$k$) and architectures (MLPs, RNNs, SSMs), Temporal Range (i) remains small in fully observed control, (ii) scales with the task's ground-truth lag in Copy-$k$, and (iii) aligns with the minimum history window required for near-optimal return as confirmed by window ablations. We also report Temporal Range for a compact Long Expressive Memory (LEM) policy trained on the task, using it as a proxy readout of task-level memory. Our axiomatic treatment draws on recent work on range measures, specialized here to temporal lag and extended to vector-valued outputs in the RL setting. Temporal Range thus offers a practical per-sequence readout of memory dependence for comparing agents and environments and for selecting the shortest sufficient context.

Paperid: 912, https://arxiv.org/pdf/2512.06102.pdf

Abstract:
Artificial intelligence methods are increasingly being explored for managing wildfires and other natural hazards. In particular, reinforcement learning (RL) is a promising path towards improving outcomes in such uncertain decision-making scenarios and moving beyond reactive strategies. However, training RL agents requires many environment interactions, and the speed of existing wildfire simulators is a severely limiting factor. We introduce $\texttt{JaxWildfire}$, a simulator underpinned by a principled probabilistic fire spread model based on cellular automata. It is implemented in JAX and enables vectorized simulations using $\texttt{vmap}$, allowing high throughput of simulations on GPUs. We demonstrate that $\texttt{JaxWildfire}$ achieves 6-35x speedup over existing software and enables gradient-based optimization of simulator parameters. Furthermore, we show that $\texttt{JaxWildfire}$ can be used to train RL agents to learn wildfire suppression policies. Our work is an important step towards enabling the advancement of RL techniques for managing natural hazards.

Paperid: 913, https://arxiv.org/pdf/2512.05994.pdf

Abstract:
With the rapid advancement of conversational and diffusion-based AI, there is a growing adoption of AI in educational services, ranging from grading and assessment tools to personalized learning systems that provide targeted support for students. However, this adaptability has yet to fully extend to the domain of children's speech, where existing models often fail due to their reliance on datasets designed for clear, articulate adult speech. Children, particularly those in early developmental stages or with speech and language pathologies, present unique challenges that current AI models and datasets are ill-equipped to handle. To address this, we introduce KidSpeak, a multi-task speech-enhanced Foundation Model capable of both generative and discriminative tasks specifically tailored to children's speech patterns. Our framework employs a two-stage training process that incorporates phonetic knowledge into the speech encoder, achieving an average accuracy of 87% across four separate tasks. Furthermore, recognizing the limitations of scalable human annotation and existing speech alignment tools, we propose the Flexible and Automatic Speech Aligner (FASA) and leverage the method to construct high quality datasets for training and evaluation. This novel alignment tool significantly improves the quality of aligned children's speech from noisy data, enhancing data quality by 13.6x compared to human annotations, as demonstrated on the CHILDES dataset. To the best of our knowledge, KidSpeak and FASA represent the first comprehensive solution designed for speech and language therapy in children, offering both a multi-purpose speech LLM and a robust alignment tool.

Paperid: 914, https://arxiv.org/pdf/2512.05925.pdf

Abstract:
How many mistakes do published AI papers contain? Peer-reviewed publications form the foundation upon which new research and knowledge are built. Errors that persist in the literature can propagate unnoticed, creating confusion in follow-up studies and complicating reproducibility. The accelerating pace of research and the increasing demands on the peer-review system make such mistakes harder to detect and avoid. To address this, we developed a Paper Correctness Checker based on GPT-5 to systematically identify mistakes in papers previously published at top AI conferences and journals. Our analysis focuses on objective mistakes-e.g., errors in formulas, derivations, calculations, figures, and tables-that have a clearly verifiable ground truth. We intentionally exclude subjective considerations such as novelty, importance, or writing quality. We find that published papers contain a non-negligible number of objective mistakes and that the average number of mistakes per paper has increased over time-from 3.8 in NeurIPS 2021 to 5.9 in NeurIPS 2025 (55.3% increase); from 4.1 in ICLR 2018 to 5.2 in ICLR 2025; and from 5.0 in TMLR 2022/23 to 5.5 in TMLR 2025. Human experts reviewed 316 potential mistakes identified by the AI Checker and confirmed that 263 were actual mistakes, corresponding to a precision of 83.2%. While most identified issues are relatively minor, correcting them would reduce confusion in the literature and strengthen reproducibility. The AI Checker also surfaced potentially more substantive mistakes that could affect the interpretation of results. Moreover, we show that the AI Checker can propose correct fixes for 75.8% of the identified mistakes. Overall, this study highlights the potential of frontier LLMs to detect and correct objective mistakes in published papers, helping to establish a firmer foundation of knowledge.

Paperid: 915, https://arxiv.org/pdf/2512.04998.pdf

Abstract:
Current upper limb prostheses aim to enhance user independence in daily activities by incorporating basic motor functions. However, they fall short of replicating the natural movement and interaction capabilities of the human arm. In contrast, human limbs leverage intrinsic compliance and actively modulate joint stiffness, enabling adaptive responses to varying tasks, impact absorption, and efficient energy transfer during dynamic actions. Inspired by this adaptability, we developed a transhumeral prosthesis with Variable Stiffness Actuators (VSAs) to replicate the controllable compliance found in biological joints. The proposed prosthesis features a modular design, allowing customization for different residual limb shapes and accommodating a range of independent control signals derived from users' biological cues. Integrated elastic elements passively support more natural movements, facilitate safe interactions with the environment, and adapt to diverse task requirements. This paper presents a comprehensive overview of the platform and its functionalities, highlighting its potential applications in the field of prosthetics.

Paperid: 916, https://arxiv.org/pdf/2512.04973.pdf

Abstract:
Variable Stiffness Actuators prove invaluable for robotics applications in unstructured environments, fostering safe interactions and enhancing task adaptability. Nevertheless, their mechanical design inevitably results in larger and heavier structures compared to classical rigid actuators. This paper introduces a novel 3 Degrees of Freedom (DoFs) parallel wrist that achieves variable stiffness through redundant elastic actuation. Leveraging its parallel architecture, the device employs only four motors, rendering it compact and lightweight. This characteristic makes it particularly well-suited for applications in prosthetics or humanoid robotics. The manuscript delves into the theoretical model of the device and proposes a sophisticated control strategy for independent regulation of joint position and stiffness. Furthermore, it validates the proposed controller through simulation, utilizing a comprehensive analysis of the system dynamics. The reported results affirm the ability of the device to achieve high accuracy and disturbance rejection in rigid configurations while minimizing interaction forces with its compliant behavior.

Paperid: 917, https://arxiv.org/pdf/2512.04875.pdf

Abstract:
Automated lesion detection in chest X-rays has demonstrated significant potential for improving clinical diagnosis by precisely localizing pathological abnormalities. While recent promptable detection frameworks have achieved remarkable accuracy in target localization, existing methods typically rely on manual annotations as prompts, which are labor-intensive and impractical for clinical applications. To address this limitation, we propose SP-Det, a novel self-prompted detection framework that automatically generates rich textual context to guide multi-label lesion detection without requiring expert annotations. Specifically, we introduce an expert-free dual-text prompt generator (DTPG) that leverages two complementary textual modalities: semantic context prompts that capture global pathological patterns and disease beacon prompts that focus on disease-specific manifestations. Moreover, we devise a bidirectional feature enhancer (BFE) that synergistically integrates comprehensive diagnostic context with disease-specific embeddings to significantly improve feature representation and detection accuracy. Extensive experiments on two chest X-ray datasets with diverse thoracic disease categories demonstrate that our SP-Det framework outperforms state-of-the-art detection methods while completely eliminating the dependency on expert-annotated prompts compared to existing promptable architectures.

Paperid: 918, https://arxiv.org/pdf/2512.04630.pdf

Abstract:
Generative AI tools, such as AI-generated hints, are increasingly integrated into programming education to offer timely, personalized support. However, little is known about how to effectively leverage these hints while ensuring autonomous and meaningful learning. One promising approach involves pairing AI-generated hints with reflection prompts, asking students to review and analyze their learning, when they request hints. This study investigates the interplay between AI-generated hints and different designs of reflection prompts in an online introductory programming course. We conducted a two-trial field experiment. In Trial 1, students were randomly assigned to receive prompts either before or after receiving hints, or no prompt at all. Each prompt also targeted one of three SRL phases: planning, monitoring, and evaluation. In Trial 2, we examined two types of prompt guidance: directed (offering more explicit and structured guidance) and open (offering more general and less constrained guidance). Findings show that students in the before-hint (RQ1), planning (RQ2), and directed (RQ3) prompt groups produced higher-quality reflections but reported lower satisfaction with AI-generated hints than those in other conditions. Immediate performance did not differ across conditions. This negative relationship between reflection quality and hint satisfaction aligns with previous work on student mental effort and satisfaction. Our results highlight the need to reconsider how AI models are trained and evaluated for education, as prioritizing user satisfaction can undermine deeper learning.

Paperid: 919, https://arxiv.org/pdf/2512.04483.pdf

Abstract:
This paper presents DeRA, a novel 1D video tokenizer that decouples the spatial-temporal representation learning in video tokenization to achieve better training efficiency and performance. Specifically, DeRA maintains a compact 1D latent space while factorizing video encoding into appearance and motion streams, which are aligned with pretrained vision foundation models to capture the spatial semantics and temporal dynamics in videos separately. To address the gradient conflicts introduced by the heterogeneous supervision, we further propose the Symmetric Alignment-Conflict Projection (SACP) module that proactively reformulates gradients by suppressing the components along conflicting directions. Extensive experiments demonstrate that DeRA outperforms LARP, the previous state-of-the-art video tokenizer by 25% on UCF-101 in terms of rFVD. Moreover, using DeRA for autoregressive video generation, we also achieve new state-of-the-art results on both UCF-101 class-conditional generation and K600 frame prediction.

Paperid: 920, https://arxiv.org/pdf/2512.04189.pdf

Abstract:
Binary Neural Networks (BNNs), which constrain both weights and activations to binary values, offer substantial reductions in computational complexity, memory footprint, and energy consumption. These advantages make them particularly well suited for deployment on resource-constrained devices. However, training BNNs via gradient-based optimization remains challenging due to the discrete nature of their variables. The dominant approach, quantization-aware training, circumvents this issue by employing surrogate gradients. Yet, this method requires maintaining latent full-precision parameters and performing the backward pass with floating-point arithmetic, thereby forfeiting the efficiency of binary operations during training. While alternative approaches based on local learning rules exist, they are unsuitable for global credit assignment and for back-propagating errors in multi-layer architectures. This paper introduces Binary Error Propagation (BEP), the first learning algorithm to establish a principled, discrete analog of the backpropagation chain rule. This mechanism enables error signals, represented as binary vectors, to be propagated backward through multiple layers of a neural network. BEP operates entirely on binary variables, with all forward and backward computations performed using only bitwise operations. Crucially, this makes BEP the first solution to enable end-to-end binary training for recurrent neural network architectures. We validate the effectiveness of BEP on both multi-layer perceptrons and recurrent neural networks, demonstrating gains of up to +6.89% and +10.57% in test accuracy, respectively. The proposed algorithm is released as an open-source repository.

Paperid: 921, https://arxiv.org/pdf/2512.03743.pdf

Abstract:
Dexterous manipulation is limited by both control and design, without consensus as to what makes manipulators best for performing dexterous tasks. This raises a fundamental challenge: how should we design and control robot manipulators that are optimized for dexterity? We present a co-design framework that learns task-specific hand morphology and complementary dexterous control policies. The framework supports 1) an expansive morphology search space including joint, finger, and palm generation, 2) scalable evaluation across the wide design space via morphology-conditioned cross-embodied control, and 3) real-world fabrication with accessible components. We evaluate the approach across multiple dexterous tasks, including in-hand rotation with simulation and real deployment. Our framework enables an end-to-end pipeline that can design, train, fabricate, and deploy a new robotic hand in under 24 hours. The full framework will be open-sourced and available on our website.

Paperid: 922, https://arxiv.org/pdf/2512.03697.pdf

Abstract:
This paper discusses the challenges encountered when analyzing the energy efficiency of synthetic benchmarks and the Gromacs package on the Fritz and Alex HPC clusters. Experiments were conducted using MPI parallelism on full sockets of Intel Ice Lake and Sapphire Rapids CPUs, as well as Nvidia A40 and A100 GPUs. The metrics and measurements obtained with the Likwid and Nvidia profiling tools are presented, along with the results. The challenges and pitfalls encountered during experimentation and analysis are revealed and discussed. Best practices for future energy efficiency analysis studies are suggested.

Paperid: 923, https://arxiv.org/pdf/2512.03538.pdf

Abstract:
World Foundation Models (WFMs) offer remarkable visual dynamics simulation capabilities, yet their application to precise robotic control remains limited by the gap between generative realism and control-oriented precision. While existing approaches use WFMs as synthetic data generators, they suffer from high computational costs and underutilization of pre-trained VLA policies. We introduce \textbf{AdaPower} (\textbf{Ada}pt and Em\textbf{power}), a lightweight adaptation framework that transforms general-purpose WFMs into specialist world models through two novel components: Temporal-Spatial Test-Time Training (TS-TTT) for inference-time adaptation and Memory Persistence (MP) for long-horizon consistency. Integrated within a Model Predictive Control framework, our adapted world model empowers pre-trained VLAs, achieving over 41\% improvement in task success rates on LIBERO benchmarks without policy retraining, while preserving computational efficiency and generalist capabilities.

Paperid: 924, https://arxiv.org/pdf/2512.03020.pdf

Abstract:
Magnetic Resonance Imaging (MRI) offers excellent soft-tissue contrast without ionizing radiation, but its long acquisition time limits clinical utility. Recent methods accelerate MRI by under-sampling $k$-space and reconstructing the resulting images using deep learning. Unrolled networks have been widely used for the reconstruction task due to their efficiency, but suffer from unstable evolving caused by freely-learnable parameters in intermediate steps. In contrast, diffusion models based on stochastic differential equations offer theoretical stability in both medical and natural image tasks but are computationally expensive. In this work, we introduce flow ODEs to MRI reconstruction by theoretically proving that unrolled networks are discrete implementations of conditional probability flow ODEs. This connection provides explicit formulations for parameters and clarifies how intermediate states should evolve. Building on this insight, we propose Flow-Aligned Training (FLAT), which derives unrolled parameters from the ODE discretization and aligns intermediate reconstructions with the ideal ODE trajectory to improve stability and convergence. Experiments on three MRI datasets show that FLAT achieves high-quality reconstructions with up to $3\times$ fewer iterations than diffusion-based generative models and significantly greater stability than unrolled networks.

Paperid: 925, https://arxiv.org/pdf/2512.02600.pdf

Abstract:
Recent years have shown increased cyber attacks targeting less secure elements in the software supply chain and causing significant damage to businesses and organizations. The US and EU governments and industry are equally interested in enhancing software security, including supply chain and vulnerability response. On June 26, 2025, researchers from the NSF-supported Secure Software Supply Chain Center (S3C2) and the Software Innovation Campus Paderborn (SICP) conducted a Vulnerability Response Summit with a diverse set of 9 practitioners from 9 companies. The goal of the Summit is to enable sharing between industry practitioners having practical experiences and challenges with software supply chain security, including vulnerability response, and helping to form new collaborations. We conducted five panel discussions based on open-ended questions regarding experiences with vulnerability reports, tools used for vulnerability discovery and management, organizational structures to report vulnerability response and management, preparedness and implementations for Cyber Resilience Act1 (CRA) and NIS22, and bug bounties. The open discussions enabled mutual sharing and shed light on common challenges that industry practitioners with practical experience face when securing their software supply chain, including vulnerability response. In this paper, we provide a summary of the Summit. Full panel questions can be found in the appendix.

Paperid: 926, https://arxiv.org/pdf/2512.02044.pdf

Abstract:
Diffusion Language Models (DLMs) have recently achieved significant success due to their any-order generation capabilities. However, existing inference methods typically rely on local, immediate-step metrics such as confidence or entropy which inherently lack a more reliable perspective. This limitation frequently leads to inconsistent sampling trajectories and suboptimal generation quality. To address this, we propose Coherent Contextual Decoding (CCD), a novel inference framework built upon two core innovations. First, CCD employs a trajectory rectification mechanism that leverages historical context to enhance sequence coherence, enabling the early rejection of suboptimal paths. We demonstrate that this mechanism is theoretically equivalent to modeling the consistency of historical steps via the conditional mutual information between context and token predictions. Building on this theoretical insight, we further address the inefficiency of conventional uniform decoding budgets. Instead of rigid allocations based on diffusion steps, we introduce an adaptive sampling strategy that dynamically adjusts the unmasking budget for each step according to our consistency metric. Consequently, our method significantly improves the quality of generation trajectories while accelerating the sampling process. Empirically, our method achieves a simultaneous enhancement in both inference speed and performance across diverse benchmarks on Dream and LLaDA, delivering up to 3.48x speedup alongside 3.91% performance improvement.

Paperid: 927, https://arxiv.org/pdf/2512.02024.pdf

Abstract:
The rapid advancement of large language models (LLMs) has accelerated their integration into clinical decision support, particularly in prescription review. To enable systematic and fine-grained evaluation, we developed RxBench, a comprehensive benchmark that covers common prescription review categories and consolidates 14 frequent types of prescription errors drawn from authoritative pharmacy references. RxBench consists of 1,150 single-choice, 230 multiple-choice, and 879 short-answer items, all reviewed by experienced clinical pharmacists. We benchmarked 18 state-of-the-art LLMs and identified clear stratification of performance across tasks. Notably, Gemini-2.5-pro-preview-05-06, Grok-4-0709, and DeepSeek-R1-0528 consistently formed the first tier, outperforming other models in both accuracy and robustness. Comparisons with licensed pharmacists indicated that leading LLMs can match or exceed human performance in certain tasks. Furthermore, building on insights from our benchmark evaluation, we performed targeted fine-tuning on a mid-tier model, resulting in a specialized model that rivals leading general-purpose LLMs in performance on short-answer question tasks. The main contribution of RxBench lies in establishing a standardized, error-type-oriented framework that not only reveals the capabilities and limitations of frontier LLMs in prescription review but also provides a foundational resource for building more reliable and specialized clinical tools.

Paperid: 928, https://arxiv.org/pdf/2512.01992.pdf

Abstract:
We introduce LLM CHESS, an evaluation framework designed to probe the generalization of reasoning and instruction-following abilities in large language models (LLMs) through extended agentic interaction in the domain of chess. We rank over 50 open and closed source models by playing against a random opponent using a range of behavioral metrics, including win and loss rates, move quality, move legality, hallucinated actions, and game duration. For a subset of top reasoning models, we derive an Elo estimate by playing against a chess engine with variably configured skill, which allows for comparisons between models in an easily understandable way. Despite the simplicity of the instruction-following task and the weakness of the opponent, many state-of-the-art models struggle to complete games or achieve consistent wins. Similar to other benchmarks on complex reasoning tasks, our experiments reveal a clear separation between reasoning and non-reasoning models. However, unlike existing static benchmarks, the stochastic and dynamic nature of LLM CHESS uniquely reduces overfitting and memorization while preventing benchmark saturation, proving difficult even for top reasoning models. To support future work on evaluating reasoning and instruction-following in LLMs, we release our experimental framework, a public leaderboard, and a dataset of associated games.

Paperid: 929, https://arxiv.org/pdf/2512.01774.pdf

Abstract:
The Segmentation Anything Model 2 (SAM2) has proven to be a powerful foundation model for promptable visual object segmentation in both images and videos, capable of storing object-aware memories and transferring them temporally through memory blocks. While SAM2 excels in video object segmentation by providing dense segmentation masks based on prompts, extending it to dense Video Semantic Segmentation (VSS) poses challenges due to the need for spatial accuracy, temporal consistency, and the ability to track multiple objects with complex boundaries and varying scales. This paper explores the extension of SAM2 for VSS, focusing on two primary approaches and highlighting firsthand observations and common challenges faced during this process. The first approach involves using SAM2 to extract unique objects as masks from a given image, with a segmentation network employed in parallel to generate and refine initial predictions. The second approach utilizes the predicted masks to extract unique feature vectors, which are then fed into a simple network for classification. The resulting classifications and masks are subsequently combined to produce the final segmentation. Our experiments suggest that leveraging SAM2 enhances overall performance in VSS, primarily due to its precise predictions of object boundaries.

Paperid: 930, https://arxiv.org/pdf/2512.01644.pdf

Abstract:
This work presents a systematic characterization of Large Language Model (LLM) inference to address fragmented understanding. Through comprehensive experiments, we establish a four-dimensional analytical framework: (1) Two-Phase Heterogeneity Observation; (2) Microarchitectural Root Cause Analysis; (3) System Scaling Principles; and (4) Emerging Paradigm Boundaries. Our investigation progresses systematically from observation to foresight: identifying performance phenomena, revealing hardware causes, validating system behavior, and exploring new paradigms. This study not only consolidates a reliable empirical foundation for existing research but also provides new discoveries and practical optimization guidance for LLM inference.

Paperid: 931, https://arxiv.org/pdf/2512.01172.pdf

Abstract:
Mean-field games (MFGs) study the Nash equilibrium of systems with a continuum of interacting agents, which can be formulated as the fixed-point of optimal control problems. They provide a unified framework for a variety of applications, including optimal transport (OT) and generative models. Despite their broad applicability, solving high-dimensional MFGs remains a significant challenge due to fundamental computational and analytical obstacles. In this work, we propose a particle-based deep Flow Matching (FM) method to tackle high-dimensional MFG computation. In each iteration of our proximal fixed-point scheme, particles are updated using first-order information, and a flow neural network is trained to match the velocity of the sample trajectories in a simulation-free manner. Theoretically, in the optimal control setting, we prove that our scheme converges to a stationary point sublinearly, and upgrade to linear (exponential) convergence under additional convexity assumptions. Our proof uses FM to induce an Eulerian coordinate (density-based) from a Lagrangian one (particle-based), and this also leads to certain equivalence results between the two formulations for MFGs when the Eulerian solution is sufficiently regular. Our method demonstrates promising performance on non-potential MFGs and high-dimensional OT problems cast as MFGs through a relaxed terminal-cost formulation.

Paperid: 932, https://arxiv.org/pdf/2512.00925.pdf

Abstract:
Accurate Multivariate Time Series (MTS) forecasting is crucial for collaborative design of complex systems, Digital Twin building, and maintenance ahead of time. However, the collaborative industrial environment presents new challenges for MTS forecasting models: models should decouple complex inter-variable dependencies while addressing non-stationary distribution shift brought by environmental changes. To address these challenges and improve collaborative sensing reliability, we propose a Patch-Based Dual-Branch Channel-Temporal Forecasting Network (D-CTNet). Particularly, with a parallel dual-branch design incorporating linear temporal modeling layer and channel attention mechanism, our method explicitly decouples and jointly learns intra-channel temporal evolution patterns and dynamic multivariate correlations. Furthermore, a global patch attention fusion module goes beyond the local window scope to model long range dependencies. Most importantly, aiming at non-stationarity, a Frequency-Domain Stationarity Correction mechanism adaptively suppresses distribution shift impacts from environment change by spectrum alignment. Evaluations on seven benchmark datasets show that our model achieves better forecasting accuracy and robustness compared with state-of-the-art methods. Our work shows great promise as a new forecasting engine for industrial collaborative systems.

Paperid: 933, https://arxiv.org/pdf/2512.00742.pdf

Abstract:
AI agents that take actions in their environment autonomously over extended time horizons require robust governance interventions to curb their potentially consequential risks. Prior proposals for governing AI agents primarily target system-level safeguards (e.g., prompt injection monitors) or agent infrastructure (e.g., agent IDs). In this work, we explore a complementary approach: regulating user interfaces of AI agents as a way of enforcing transparency and behavioral requirements that then demand changes at the system and/or infrastructure levels. Specifically, we analyze 22 existing agentic systems to identify UI elements that play key roles in human-agent interaction and communication. We then synthesize those elements into six high-level interaction design patterns that hold regulatory potential (e.g., requiring agent memory to be editable). We conclude with policy recommendations based on our analysis. Our work exposes a new surface for regulatory action that supplements previous proposals for practical AI agent governance.

Paperid: 934, https://arxiv.org/pdf/2512.00534.pdf

Abstract:
Maintaining consistent 3D scene representations over time is a significant challenge in computer vision. Updating 3D scenes from sparse-view observations is crucial for various real-world applications, including urban planning, disaster assessment, and historical site preservation, where dense scans are often unavailable or impractical. In this paper, we propose Cross-Temporal 3D Gaussian Splatting (Cross-Temporal 3DGS), a novel framework for efficiently reconstructing and updating 3D scenes across different time periods, using sparse images and previously captured scene priors. Our approach comprises three stages: 1) Cross-temporal camera alignment for estimating and aligning camera poses across different timestamps; 2) Interference-based confidence initialization to identify unchanged regions between timestamps, thereby guiding updates; and 3) Progressive cross-temporal optimization, which iteratively integrates historical prior information into the 3D scene to enhance reconstruction quality. Our method supports non-continuous capture, enabling not only updates using new sparse views to refine existing scenes, but also recovering past scenes from limited data with the help of current captures. Furthermore, we demonstrate the potential of this approach to achieve temporal changes using only sparse images, which can later be reconstructed into detailed 3D representations as needed. Experimental results show significant improvements over baseline methods in reconstruction quality and data efficiency, making this approach a promising solution for scene versioning, cross-temporal digital twins, and long-term spatial documentation.

Paperid: 935, https://arxiv.org/pdf/2512.00135.pdf

Abstract:
We study the statistical design of a fair mechanism that attains equalized odds, where an agent uses some useful data (database) $X$ to solve a task $T$. Since both $X$ and $T$ are correlated with some latent sensitive attribute $S$, the agent designs a representation $Y$ that satisfies an equalized odds, that is, such that $I(Y;S|T) =0$. In contrast to our previous work, we assume here that the agent has no direct access to $S$ and $T$; hence, the Markov chains $S - X - Y$ and $T - X - Y$ hold. Furthermore, we impose a geometric structure on the conditional distribution $P_{S|Y}$, allowing $Y$ and $S$ to have a small correlation, bounded by a threshold. When the threshold is small, concepts from information geometry allow us to approximate mutual information and reformulate the fair mechanism design problem as a quadratic program with closed-form solutions under certain constraints. For other cases, we derive simple, low-complexity lower bounds based on the maximum singular value and vector of a matrix. Finally, we compare our designs with the optimal solution in a numerical example.

Paperid: 936, https://arxiv.org/pdf/2512.07673.pdf

Abstract:
Effective motion representation is crucial for enabling robots to imitate expressive behaviors in real time, yet existing motion controllers often ignore inherent patterns in motion. Previous efforts in representation learning do not attempt to jointly capture structured periodic patterns and irregular variations in human and animal movement. To address this, we present Multi-Domain Motion Embedding (MDME), a motion representation that unifies the embedding of structured and unstructured features using a wavelet-based encoder and a probabilistic embedding in parallel. This produces a rich representation of reference motions from a minimal input set, enabling improved generalization across diverse motion styles and morphologies. We evaluate MDME on retargeting-free real-time motion imitation by conditioning robot control policies on the learned embeddings, demonstrating accurate reproduction of complex trajectories on both humanoid and quadruped platforms. Our comparative studies confirm that MDME outperforms prior approaches in reconstruction fidelity and generalizability to unseen motions. Furthermore, we demonstrate that MDME can reproduce novel motion styles in real-time through zero-shot deployment, eliminating the need for task-specific tuning or online retargeting. These results position MDME as a generalizable and structure-aware foundation for scalable real-time robot imitation.

Paperid: 937, https://arxiv.org/pdf/2512.07638.pdf

Abstract:
The emergence of sixth-generation (6G) networks marks a paradigm shift: by unifying an edge-to-cloud computing continuum with ultra-high-performance networking, 6G will enable capabilities far beyond today's boundaries. As use-case diversity grows exponentially and user adoption drives traffic to unprecedented and highly dynamic levels, novel service orchestration mechanisms are indispensable. In this paper, we adopt an architectural viewpoint, examining Service Registration, Indexing, Discovery, and Selection (SRIDS) as fundamental elements of 6G service provision. We first establish the theoretical foundations of SRIDS in 6G by defining its core concepts, detailing its end-to-end workflow, reviewing current standardization efforts, and projecting its future design objectives, including reliability, scalability, automaticity and adaptability, determinism, efficiency, sustainability, semantic-awareness, security, privacy, and trust. We then perform a comprehensive literature review and gap analysis encompassing both existing surveys and recent research efforts, identifying conceptual and methodological gaps that hinder unified SRIDS in 6G. Next, we introduce a taxonomy that classifies SRIDS mechanisms into centralized, distributed, decentralized, and hybrid architectures, and systematically examine the relevant studies within each category. Each work is evaluated against the extracted design objectives. Building on these findings, we propose a hybrid architectural framework, combining centralized data management to ensure consistency and agility with distributed coordination to enhance scalability in emerging 6G use cases. The framework incorporates innovative technologies, such as Generative Artificial Intelligence (GenAI). We conclude by highlighting open challenges and suggesting directions for future research.

Paperid: 938, https://arxiv.org/pdf/2512.07578.pdf

Abstract:
We propose $ϕ$-test, a global feature-selection and significance procedure for black-box predictors that combines Shapley attributions with selective inference. Given a trained model and an evaluation dataset, $ϕ$-test performs SHAP-guided screening and fits a linear surrogate on the screened features via a selection rule with a tractable selective-inference form. For each retained feature, it outputs a Shapley-based global score, a surrogate coefficient, and post-selection $p$-values and confidence intervals in a global feature-importance table. Experiments on real tabular regression tasks with tree-based and neural backbones suggest that $ϕ$-test can retain much of the predictive ability of the original model while using only a few features and producing feature sets that remain fairly stable across resamples and backbone classes. In these settings, $ϕ$-test acts as a practical global explanation layer linking Shapley-based importance summaries with classical statistical inference.

Paperid: 939, https://arxiv.org/pdf/2512.07536.pdf

Abstract:
Network topology is critical for efficient parameter synchronization in distributed learning over networks. However, most existing studies do not account for bandwidth limitations in network topology design. In this paper, we propose a bandwidth-aware network topology optimization framework to maximize consensus speed under edge cardinality constraints. For heterogeneous bandwidth scenarios, we introduce a maximum bandwidth allocation strategy for the edges to ensure efficient communication among nodes. By reformulating the problem into an equivalent Mixed-Integer SDP problem, we leverage a computationally efficient ADMM-based method to obtain topologies that yield the maximum consensus speed. Within the ADMM substep, we adopt the conjugate gradient method to efficiently solve large-scale linear equations to achieve better scalability. Experimental results demonstrate that the resulting network topologies outperform the benchmark topologies in terms of consensus speed, and reduce the training time required for decentralized learning tasks on real-world datasets to achieve the target test accuracy, exhibiting speedups of more than $1.11\times$ and $1.21\times$ for homogeneous and heterogeneous bandwidth settings, respectively.

Paperid: 940, https://arxiv.org/pdf/2512.07195.pdf

Abstract:
Multi-agent role-playing has recently shown promise for studying social behavior with language agents, but existing simulations are mostly monolingual and fail to model cross-lingual interaction, an essential property of real societies. We introduce MASim, the first multilingual agent-based simulation framework that supports multi-turn interaction among generative agents with diverse sociolinguistic profiles. MASim offers two key analyses: (i) global public opinion modeling, by simulating how attitudes toward open-domain hypotheses evolve across languages and cultures, and (ii) media influence and information diffusion, via autonomous news agents that dynamically generate content and shape user behavior. To instantiate simulations, we construct the MAPS benchmark, which combines survey questions and demographic personas drawn from global population distributions. Experiments on calibration, sensitivity, consistency, and cultural case studies show that MASim reproduces sociocultural phenomena and highlights the importance of multilingual simulation for scalable, controlled computational social science.

Paperid: 941, https://arxiv.org/pdf/2512.06987.pdf

Abstract:
Accurately predicting experimentally-realizable 3D molecular crystal structures from their 2D chemical graphs is a long-standing open challenge in computational chemistry called crystal structure prediction (CSP). Efficiently solving this problem has implications ranging from pharmaceuticals to organic semiconductors, as crystal packing directly governs the physical and chemical properties of organic solids. In this paper, we introduce OXtal, a large-scale 100M parameter all-atom diffusion model that directly learns the conditional joint distribution over intramolecular conformations and periodic packing. To efficiently scale OXtal, we abandon explicit equivariant architectures imposing inductive bias arising from crystal symmetries in favor of data augmentation strategies. We further propose a novel crystallization-inspired lattice-free training scheme, Stoichiometric Stochastic Shell Sampling ($S^4$), that efficiently captures long-range interactions while sidestepping explicit lattice parametrization -- thus enabling more scalable architectural choices at all-atom resolution. By leveraging a large dataset of 600K experimentally validated crystal structures (including rigid and flexible molecules, co-crystals, and solvates), OXtal achieves orders-of-magnitude improvements over prior ab initio machine learning CSP methods, while remaining orders of magnitude cheaper than traditional quantum-chemical approaches. Specifically, OXtal recovers experimental structures with conformer $\text{RMSD}_1<0.5$ Å and attains over 80\% packing similarity rate, demonstrating its ability to model both thermodynamic and kinetic regularities of molecular crystallization.

Paperid: 942, https://arxiv.org/pdf/2512.06849.pdf

Abstract:
Accurate segmentation of vertebral metastasis in CT is clinically important yet difficult to scale, as voxel-level annotations are scarce and both lytic and blastic lesions often resemble benign degenerative changes. We introduce a weakly supervised method trained solely on vertebra-level healthy/malignant labels, without any lesion masks. The method combines a Diffusion Autoencoder (DAE) that produces a classifier-guided healthy edit of each vertebra with pixel-wise difference maps that propose candidate lesion regions. To determine which regions truly reflect malignancy, we introduce Hide-and-Seek Attribution: each candidate is revealed in turn while all others are hidden, the edited image is projected back to the data manifold by the DAE, and a latent-space classifier quantifies the isolated malignant contribution of that component. High-scoring regions form the final lytic or blastic segmentation. On held-out radiologist annotations, we achieve strong blastic/lytic performance despite no mask supervision (F1: 0.91/0.85; Dice: 0.87/0.78), exceeding baselines (F1: 0.79/0.67; Dice: 0.74/0.55). These results show that vertebra-level labels can be transformed into reliable lesion masks, demonstrating that generative editing combined with selective occlusion supports accurate weakly supervised segmentation in CT.

Paperid: 943, https://arxiv.org/pdf/2512.06700.pdf

Abstract:
Live-streaming, as an emerging media enabling real-time interaction between authors and users, has attracted significant attention. Unlike the stable playback time of traditional TV live or the fixed content of short video, live-streaming, due to the dynamics of content and time, poses higher requirements for the recommendation algorithm of the platform - understanding the ever-changing content in real time and push it to users at the appropriate moment. Through analysis, we find that users have a better experience and express more positive behaviors during highlight moments of the live-streaming. Furthermore, since the model lacks access to future content during recommendation, yet user engagement depends on how well subsequent content aligns with their interests, an intuitive solution is to predict future live-streaming content. Therefore, we perform semantic quantization on live-streaming segments to obtain Semantic ids (Sid), encode the historical Sid sequence to capture the author's characteristics, and model Sid evolution trend to enable foresight prediction of future content. This foresight enhances the ranking model through refined features. Extensive offline and online experiments demonstrate the effectiveness of our method.

Paperid: 944, https://arxiv.org/pdf/2512.06678.pdf

Abstract:
Instruction tuning is one of the key steps required for adapting large language models (LLMs) to a broad spectrum of downstream applications. However, this procedure is difficult because real-world datasets are rarely homogeneous; they consist of a mixture of diverse information, causing gradient interference, where conflicting gradients pull the model in opposing directions, degrading performance. A common strategy to mitigate this issue is to group data based on semantic or embedding similarity. However, this fails to capture how data influences model parameters during learning. While recent works have attempted to cluster gradients directly, they randomly project gradients into lower dimensions to manage memory, which leads to accuracy loss. Moreover, these methods rely on expert ensembles which necessitates multiple inference passes and expensive on-the-fly gradient computations during inference. To address these limitations, we propose GradientSpace, a framework that clusters samples directly in full-dimensional gradient space. We introduce an online SVD-based algorithm that operates on LoRA gradients to identify latent skills without the infeasible cost of storing all sample gradients. Each cluster is used to train a specialized LoRA expert along with a lightweight router trained to select the best expert during inference. We show that routing to a single, appropriate expert outperforms expert ensembles used in prior work, while significantly reducing inference latency. Our experiments across mathematical reasoning, code generation, finance, and creative writing tasks demonstrate that GradientSpace leads to coherent expert specialization and consistent accuracy gains over state-of-the-art clustering methods and finetuning techniques.

Paperid: 945, https://arxiv.org/pdf/2512.05501.pdf

Abstract:
Safeguard models help large language models (LLMs) detect and block harmful content, but most evaluations remain English-centric and overlook linguistic and cultural diversity. Existing multilingual safety benchmarks often rely on machine-translated English data, which fails to capture nuances in low-resource languages. Southeast Asian (SEA) languages are underrepresented despite the region's linguistic diversity and unique safety concerns, from culturally sensitive political speech to region-specific misinformation. Addressing these gaps requires benchmarks that are natively authored to reflect local norms and harm scenarios. We introduce SEA-SafeguardBench, the first human-verified safety benchmark for SEA, covering eight languages, 21,640 samples, across three subsets: general, in-the-wild, and content generation. The experimental results from our benchmark demonstrate that even state-of-the-art LLMs and guardrails are challenged by SEA cultural and harm scenarios and underperform when compared to English texts.

Paperid: 946, https://arxiv.org/pdf/2512.05397.pdf

Abstract:
Major life transitions demand high-stakes decisions, yet people often struggle to imagine how their future selves will live with the consequences. To support this limited capacity for mental time travel, we introduce AI-enabled digital twins that have ``lived through'' simulated life scenarios. Rather than predicting optimal outcomes, these simulations extend prospective cognition by making alternative futures vivid enough to support deliberation without assuming which path is best. We evaluate this idea in a randomized controlled study (N=192) using multimodal synthesis - facial age progression, voice cloning, and large language model dialogue - to create personalized avatars representing participants 30 years forward. Young adults 18 to 28 years old described pending binary decisions and were assigned to guided imagination or one of four avatar conditions: single-option, balanced dual-option, or expanded three-option with a system-generated novel alternative. Results showed asymmetric effects: single-sided avatars increased shifts toward the presented option, while balanced presentation produced movement toward both. Introducing a system-generated third option increased adoption of this new alternative compared to control, suggesting that AI-generated future selves can expand choice by surfacing paths that might otherwise go unnoticed. Participants rated evaluative reasoning and eudaimonic meaning-making as more important than emotional or visual vividness. Perceived persuasiveness and baseline agency predicted decision change. These findings advance understanding of AI-mediated episodic prospection and raise questions about autonomy in AI-augmented decisions.

Paperid: 947, https://arxiv.org/pdf/2512.05128.pdf

Abstract:
Jamming devices disrupt signals from the global navigation satellite system (GNSS) and pose a significant threat by compromising the reliability of accurate positioning. Consequently, the detection and localization of these interference signals are essential to achieve situational awareness, mitigating their impact, and implementing effective countermeasures. In this paper, we utilize a two-times-two patch antenna system (i.e., the software defined radio device Ettus USRP X440) to predict the angle, elevation, and distance to the jamming source based on in-phase and quadrature (IQ) samples. We propose to use an inertial measurement unit (IMU) attached to the antenna system to predict the relative movement of the antenna in dynamic scenarios. We present a synthetic aperture system that enables coherent spatial imaging using platform motion to synthesize larger virtual apertures, offering superior angular resolution without mechanically rotating antennas. While classical angle-of-arrival (AoA) methods exhibit reduced accuracy in multipath environments due to signal reflections and scattering, leading to localization errors, we utilize a methodology that fuses IQ and Fast Fourier Transform (FFT)-computed spectrograms with 22 AoA features and the predicted relative movement to enhance GNSS jammer direction finding.

Paperid: 948, https://arxiv.org/pdf/2512.04960.pdf

Abstract:
Despite the fact that visuomotor-based policies obtained via imitation learning demonstrate good performances in complex manipulation tasks, they usually struggle to achieve the same accuracy and speed as traditional control based methods. In this work, we introduce Hybrid-Diffusion models that combine open-loop routines with visuomotor diffusion policies. We develop Teleoperation Augmentation Primitives (TAPs) that allow the operator to perform predefined routines, such as locking specific axes, moving to perching waypoints, or triggering task-specific routines seamlessly during demonstrations. Our Hybrid-Diffusion method learns to trigger such TAPs during inference. We validate the method on challenging real-world tasks: Vial Aspiration, Open-Container Liquid Transfer, and container unscrewing. All experimental videos are available on the project's website: https://hybriddiffusion.github.io/

Paperid: 949, https://arxiv.org/pdf/2512.03358.pdf

Abstract:
Quantum Federated Learning (QFL) promises to revolutionize distributed machine learning by combining the computational power of quantum devices with collaborative model training. Yet, privacy of both data and models remains a critical challenge. In this work, we propose a privacy-preserving QFL framework where a network of $n$ quantum devices trains local models and transmits them to a central server under a multi-layered privacy protocol. Our design leverages Singular Value Decomposition (SVD), Quantum Key Distribution (QKD), and Analytic Quantum Gradient Descent (AQGD) to secure data preparation, model sharing, and training stages. Through theoretical analysis and experiments on contemporary quantum platforms and datasets, we demonstrate that the framework robustly safeguards data and model confidentiality while maintaining training efficiency.

Paperid: 950, https://arxiv.org/pdf/2512.03275.pdf

Abstract:
In this paper, we study the parameterized complexity of local search, whose goal is to find a good nearby solution from the given current solution. Formally, given an optimization problem where the goal is to find the largest feasible subset $S$ of a universe $U$, the new input consists of a current solution $P$ (not necessarily feasible) as well as an ordinary input for the problem. Given the existence of a feasible solution $S^*$, the goal is to find a feasible solution as good as $S^*$ in parameterized time $f(k) \cdot n^{O(1)}$, where $k$ denotes the distance $|PΔS^*|$. This model generalizes numerous classical parameterized optimization problems whose parameter $k$ is the minimum number of elements removed from $U$ to make it feasible, which corresponds to the case $P = U$. We apply this model to widely studied Constraint Satisfaction Problems (CSPs), where $U$ is the set of constraints, and a subset $U'$ of constraints is feasible if there is an assignment to the variables satisfying all constraints in $U'$. We give a complete characterization of the parameterized complexity of all boolean-alphabet symmetric CSPs, where the predicate's acceptance depends on the number of true literals.

Paperid: 951, https://arxiv.org/pdf/2512.03262.pdf

Abstract:
Vibe coding is a new programming paradigm in which human engineers instruct large language model (LLM) agents to complete complex coding tasks with little supervision. Although it is increasingly adopted, are vibe coding outputs really safe to deploy in production? To answer this question, we propose SU S VI B E S, a benchmark consisting of 200 feature-request software engineering tasks from real-world open-source projects, which, when given to human programmers, led to vulnerable implementations. We evaluate multiple widely used coding agents with frontier models on this benchmark. Disturbingly, all agents perform poorly in terms of software security. Although 61% of the solutions from SWE-Agent with Claude 4 Sonnet are functionally correct, only 10.5% are secure. Further experiments demonstrate that preliminary security strategies, such as augmenting the feature request with vulnerability hints, cannot mitigate these security issues. Our findings raise serious concerns about the widespread adoption of vibe-coding, particularly in security-sensitive applications.

Paperid: 952, https://arxiv.org/pdf/2512.03000.pdf

Abstract:
Understanding the dynamic physical world, characterized by its evolving 3D structure, real-world motion, and semantic content with textual descriptions, is crucial for human-agent interaction and enables embodied agents to perceive and act within real environments with human-like capabilities. However, existing datasets are often derived from limited simulators or utilize traditional Structurefrom-Motion for up-to-scale annotation and offer limited descriptive captioning, which restricts the capacity of foundation models to accurately interpret real-world dynamics from monocular videos, commonly sourced from the internet. To bridge these gaps, we introduce DynamicVerse, a physical-scale, multimodal 4D world modeling framework for dynamic real-world video. We employ large vision, geometric, and multimodal models to interpret metric-scale static geometry, real-world dynamic motion, instance-level masks, and holistic descriptive captions. By integrating window-based Bundle Adjustment with global optimization, our method converts long real-world video sequences into a comprehensive 4D multimodal format. DynamicVerse delivers a large-scale dataset consisting of 100K+ videos with 800K+ annotated masks and 10M+ frames from internet videos. Experimental evaluations on three benchmark tasks, namely video depth estimation, camera pose estimation, and camera intrinsics estimation, demonstrate that our 4D modeling achieves superior performance in capturing physical-scale measurements with greater global accuracy than existing methods.

Paperid: 953, https://arxiv.org/pdf/2512.02700.pdf

Abstract:
Vision-language models (VLMs) excel at image understanding tasks, but the large number of visual tokens imposes significant computational costs, hindering deployment on mobile devices. Many pruning methods rely solely on token importance and thus overlook inter-token redundancy, retaining numerous duplicated tokens and wasting capacity. Although some redundancy-aware approaches have been proposed, they often ignore the spatial relationships among visual tokens. This can lead to overly sparse selections of retained tokens that fail to adequately cover the regions of target objects. To address these limitations, we propose VLM-Pruner, a training-free token pruning algorithm that explicitly balances redundancy and spatial sparsity. We introduce a centrifugal token pruning paradigm that enables near-to-far selection while prioritizing the preservation of fine-grained object details. Moreover, we design a Buffering for Spatial Sparsity (BSS) criterion that defers the selection of spatially distant tokens. We further adopt a parallel greedy strategy to conduct token selection efficiently. To mitigate information loss from pruning, we selectively fuse salient information from the discarded tokens into the retained ones. Comprehensive comparisons demonstrate that VLM-Pruner consistently outperforms strong baselines across five VLMs with an 88.9\% pruning rate, while delivering an end-to-end inference speedup.

Paperid: 954, https://arxiv.org/pdf/2512.02556.pdf

Authors:DeepSeek-AI, Aixin Liu, Aoxue Mei, Bangcai Lin, Bing Xue, Bingxuan Wang, Bingzheng Xu, Bochao Wu, Bowei Zhang, Chaofan Lin, Chen Dong, Chengda Lu, Chenggang Zhao, Chengqi Deng, Chenhao Xu, Chong Ruan, Damai Dai, Daya Guo, Dejian Yang, Deli Chen, Erhang Li, Fangqi Zhou, Fangyun Lin, Fucong Dai, Guangbo Hao, Guanting Chen, Guowei Li, H. Zhang, Hanwei Xu, Hao Li, Haofen Liang, Haoran Wei, Haowei Zhang, Haowen Luo, Haozhe Ji, Honghui Ding, Hongxuan Tang, Huanqi Cao, Huazuo Gao, Hui Qu, Hui Zeng, Jialiang Huang, Jiashi Li, Jiaxin Xu, Jiewen Hu, Jingchang Chen, Jingting Xiang, Jingyang Yuan, Jingyuan Cheng, Jinhua Zhu, Jun Ran, Junguang Jiang, Junjie Qiu, Junlong Li, Junxiao Song, Kai Dong, Kaige Gao, Kang Guan, Kexin Huang, Kexing Zhou, Kezhao Huang, Kuai Yu, Lean Wang, Lecong Zhang, Lei Wang, Liang Zhao, Liangsheng Yin, Lihua Guo, Lingxiao Luo, Linwang Ma, Litong Wang, Liyue Zhang, M. S. Di, M. Y Xu, Mingchuan Zhang, Minghua Zhang, Minghui Tang, Mingxu Zhou, Panpan Huang, Peixin Cong, Peiyi Wang, Qiancheng Wang, Qihao Zhu, Qingyang Li, Qinyu Chen, Qiushi Du, Ruiling Xu, Ruiqi Ge, Ruisong Zhang, Ruizhe Pan, Runji Wang, Runqiu Yin, Runxin Xu, Ruomeng Shen, Ruoyu Zhang, S. H. Liu, Shanghao Lu, Shangyan Zhou, Shanhuang Chen, Shaofei Cai, Shaoyuan Chen, Shengding Hu, Shengyu Liu, Shiqiang Hu, Shirong Ma, Shiyu Wang, Shuiping Yu, Shunfeng Zhou, Shuting Pan, Songyang Zhou, Tao Ni, Tao Yun, Tian Pei, Tian Ye, Tianyuan Yue, Wangding Zeng, Wen Liu, Wenfeng Liang, Wenjie Pang, Wenjing Luo, Wenjun Gao, Wentao Zhang, Xi Gao, Xiangwen Wang, Xiao Bi, Xiaodong Liu, Xiaohan Wang, Xiaokang Chen, Xiaokang Zhang, Xiaotao Nie, Xin Cheng, Xin Liu, Xin Xie, Xingchao Liu, Xingkai Yu, Xingyou Li, Xinyu Yang, Xinyuan Li, Xu Chen, Xuecheng Su, Xuehai Pan, Xuheng Lin, Xuwei Fu, Y. Q. Wang, Yang Zhang, Yanhong Xu, Yanru Ma, Yao Li, Yao Li, Yao Zhao, Yaofeng Sun, Yaohui Wang, Yi Qian, Yi Yu, Yichao Zhang, Yifan Ding, Yifan Shi, Yiliang Xiong, Ying He, Ying Zhou, Yinmin Zhong, Yishi Piao, Yisong Wang, Yixiao Chen, Yixuan Tan, Yixuan Wei, Yiyang Ma, Yiyuan Liu, Yonglun Yang, Yongqiang Guo, Yongtong Wu, Yu Wu, Yuan Cheng, Yuan Ou, Yuanfan Xu, Yuduan Wang, Yue Gong, Yuhan Wu, Yuheng Zou, Yukun Li, Yunfan Xiong, Yuxiang Luo, Yuxiang You, Yuxuan Liu, Yuyang Zhou, Z. F. Wu, Z. Z. Ren, Zehua Zhao, Zehui Ren, Zhangli Sha, Zhe Fu, Zhean Xu, Zhenda Xie, Zhengyan Zhang, Zhewen Hao, Zhibin Gou, Zhicheng Ma, Zhigang Yan, Zhihong Shao, Zhixian Huang, Zhiyu Wu, Zhuoshu Li, Zhuping Zhang, Zian Xu, Zihao Wang, Zihui Gu, Zijia Zhu, Zilin Li, Zipeng Zhang, Ziwei Xie, Ziyi Gao, Zizheng Pan, Zongqing Yao, Bei Feng, Hui Li, J. L. Cai, Jiaqi Ni, Lei Xu, Meng Li, Ning Tian, R. J. Chen, R. L. Jin, S. S. Li, Shuang Zhou, Tianyu Sun, X. Q. Li, Xiangyue Jin, Xiaojin Shen, Xiaosha Chen, Xinnan Song, Xinyi Zhou, Y. X. Zhu, Yanping Huang, Yaohui Li, Yi Zheng, Yuchen Zhu, Yunxian Ma, Zhen Huang, Zhipeng Xu, Zhongyu Zhang, Dongjie Ji, Jian Liang, Jianzhong Guo, Jin Chen, Leyi Xia, Miaojun Wang, Mingming Li, Peng Zhang, Ruyi Chen, Shangmian Sun, Shaoqing Wu, Shengfeng Ye, T. Wang, W. L. Xiao, Wei An, Xianzu Wang, Xiaowen Sun, Xiaoxiang Wang, Ying Tang, Yukun Zha, Zekai Zhang, Zhe Ju, Zhen Zhang, Zihua Qu

Abstract:
We introduce DeepSeek-V3.2, a model that harmonizes high computational efficiency with superior reasoning and agent performance. The key technical breakthroughs of DeepSeek-V3.2 are as follows: (1) DeepSeek Sparse Attention (DSA): We introduce DSA, an efficient attention mechanism that substantially reduces computational complexity while preserving model performance in long-context scenarios. (2) Scalable Reinforcement Learning Framework: By implementing a robust reinforcement learning protocol and scaling post-training compute, DeepSeek-V3.2 performs comparably to GPT-5. Notably, our high-compute variant, DeepSeek-V3.2-Speciale, surpasses GPT-5 and exhibits reasoning proficiency on par with Gemini-3.0-Pro, achieving gold-medal performance in both the 2025 International Mathematical Olympiad (IMO) and the International Olympiad in Informatics (IOI). (3) Large-Scale Agentic Task Synthesis Pipeline: To integrate reasoning into tool-use scenarios, we developed a novel synthesis pipeline that systematically generates training data at scale. This methodology facilitates scalable agentic post-training, yielding substantial improvements in generalization and instruction-following robustness within complex, interactive environments.

Paperid: 955, https://arxiv.org/pdf/2512.02301.pdf

Abstract:
This work presents vQFL (vehicular Quantum Federated Learning), a new framework that leverages quantum machine learning techniques to tackle key privacy and security issues in autonomous vehicular networks. Furthermore, we propose a server-side adapted fine-tuning method, ft-VQFL,to achieve enhanced and more resilient performance. By integrating quantum federated learning with differential privacy and quantum key distribution (QKD), our quantum vanguard approach creates a multi-layered defense against both classical and quantum threats while preserving model utility. Extensive experimentation with industry-standard datasets (KITTI, Waymo, and nuScenes) demonstrates that vQFL maintains accuracy comparable to standard QFL while significantly improving privacy guaranties and communication security. Our implementation using various quantum models (VQC, QCNN, and SamplerQNN) reveals minimal performance overhead despite the added security measures. This work establishes a crucial foundation for quantum-resistant autonomous vehicle systems that can operate securely in the post-quantum era while efficiently processing the massive data volumes (20-40TB/day per vehicle) generated by modern autonomous fleets. The modular design of the framework allows for seamless integration with existing vehicular networks, positioning vQFL as an essential component for future intelligent transportation infrastructure.

Paperid: 956, https://arxiv.org/pdf/2512.02061.pdf

Abstract:
Multivariate time series forecasts are widely used, such as industrial, transportation and financial forecasts. However, the dominant frequencies in time series may shift with the evolving spectral distribution of the data. Traditional Mixture of Experts (MoE) models, which employ a fixed number of experts, struggle to adapt to these changes, resulting in frequency coverage imbalance issue. Specifically, too few experts can lead to the overlooking of critical information, while too many can introduce noise. To this end, we propose Ada-MoGE, an adaptive Gaussian Mixture of Experts model. Ada-MoGE integrates spectral intensity and frequency response to adaptively determine the number of experts, ensuring alignment with the input data's frequency distribution. This approach prevents both information loss due to an insufficient number of experts and noise contamination from an excess of experts. Additionally, to prevent noise introduction from direct band truncation, we employ Gaussian band-pass filtering to smoothly decompose the frequency domain features, further optimizing the feature representation. The experimental results show that our model achieves state-of-the-art performance on six public benchmarks with only 0.2 million parameters.

Paperid: 957, https://arxiv.org/pdf/2512.01732.pdf

Abstract:
In distributed and federated learning algorithms, communication overhead is often reduced by performing multiple local updates between communication rounds. However, due to data heterogeneity across nodes and the local gradient noise within each node, this strategy can lead to the drift of local models away from the global optimum. To address this issue, we revisit the well-known federated learning method Scaffold (Karimireddy et al., 2020) under a gradient tracking perspective, and propose a unified spatio-temporal gradient tracking algorithm, termed ST-GT, for distributed stochastic optimization over time-varying graphs. ST-GT tracks the global gradient across neighboring nodes to mitigate data heterogeneity, while maintaining a running average of local gradients to substantially suppress noise, with slightly more storage overhead. Without assuming bounded data heterogeneity, we prove that ST-GT attains a linear convergence rate for strongly convex problems and a sublinear rate for nonconvex cases. Notably, ST-GT achieves the first linear speed-up in communication complexity with respect to the number of local updates per round $τ$ for the strongly-convex setting. Compared to traditional gradient tracking methods, ST-GT reduces the topology-dependent noise term from $σ^2$ to $σ^2/τ$, where $σ^2$ denotes the noise level, thereby improving communication efficiency.

Paperid: 958, https://arxiv.org/pdf/2512.01556.pdf

Abstract:
Large language models (LLMs) often generate unreliable answers, while heuristic uncertainty methods fail to fully distinguish correct from incorrect predictions, causing users to accept erroneous answers without statistical guarantees. We address this issue through the lens of false discovery rate (FDR) control, ensuring that among all accepted predictions, the proportion of errors does not exceed a target risk level. To achieve this in a principled way, we propose LEC, which reinterprets selective prediction as a constrained decision problem by enforcing a Linear Expectation Constraint over selection and error indicators. Then, we establish a finite-sample sufficient condition, which relies only on a held-out set of exchangeable calibration samples, to compute an FDR-constrained, coverage-maximizing threshold. Furthermore, we extend LEC to a two-model routing mechanism: given a prompt, if the current model's uncertainty exceeds its calibrated threshold, we delegate it to a stronger model, while maintaining a unified FDR guarantee. Evaluations on closed-ended and open-ended question-answering (QA) datasets show that LEC achieves tighter FDR control and substantially improves sample retention over prior methods. Moreover, the two-model routing mechanism achieves lower risk levels while accepting more correct samples than each individual model.

Paperid: 959, https://arxiv.org/pdf/2512.01384.pdf

Abstract:
We present CLAPS, a posterior-aware conformal regression method that pairs a Last-Layer Laplace Approximation with split-conformal calibration. From the resulting Gaussian posterior, CLAPS defines a simple two-sided posterior CDF score that aligns the conformity metric with the full predictive shape, not just a point estimate. This alignment yields narrower prediction intervals at the same target coverage, especially on small to medium tabular datasets where data are scarce and uncertainty modeling matters. We also provide a lightweight diagnostic suite that separates aleatoric and epistemic components and visualizes posterior behavior, helping practitioners understand why intervals shrink when they do. Across multiple benchmarks using the same MLP backbone, CLAPS consistently attains nominal coverage with improved efficiency and minimal overhead, offering a clear, practical upgrade to residual-based conformal baselines.

Paperid: 960, https://arxiv.org/pdf/2512.01280.pdf

Abstract:
Autonomous aerial tracking with drones offers vast potential for surveillance, cinematography, and industrial inspection applications. While single-drone tracking systems have been extensively studied, swarm-based target tracking remains underexplored, despite its unique advantages of distributed perception, fault-tolerant redundancy, and multidirectional target coverage. To bridge this gap, we propose a novel decentralized LiDAR-based swarm tracking framework that enables visibility-aware, cooperative target tracking in complex environments, while fully harnessing the unique capabilities of swarm systems. To address visibility, we introduce a novel Spherical Signed Distance Field (SSDF)-based metric for 3-D environmental occlusion representation, coupled with an efficient algorithm that enables real-time onboard SSDF updating. A general Field-of-View (FOV) alignment cost supporting heterogeneous LiDAR configurations is proposed for consistent target observation. Swarm coordination is enhanced through cooperative costs that enforce inter-robot safe clearance, prevent mutual occlusions, and notably facilitate 3-D multidirectional target encirclement via a novel electrostatic-potential-inspired distribution metric. These innovations are integrated into a hierarchical planner, combining a kinodynamic front-end searcher with a spatiotemporal $SE(3)$ back-end optimizer to generate collision-free, visibility-optimized trajectories.Deployed on heterogeneous LiDAR swarms, our fully decentralized implementation features collaborative perception, distributed planning, and dynamic swarm reconfigurability. Validated through rigorous real-world experiments in cluttered outdoor environments, the proposed system demonstrates robust cooperative tracking of agile targets (drones, humans) while achieving superior visibility maintenance.

Paperid: 961, https://arxiv.org/pdf/2512.01278.pdf

Abstract:
Reasoning language models have demonstrated remarkable capabilities on challenging tasks by generating elaborate chain-of-thought (CoT) solutions. However, such lengthy generation shifts the inference bottleneck from compute-bound to memory-bound. To generate each token, the model applies full attention to all previously generated tokens, requiring memory access to an increasingly large KV-Cache. Consequently, longer generations demand more memory access for every step, leading to substantial pressure on memory bandwidth. To address this, we introduce SparseSpec, a speculative decoding framework that reuses the same model as the draft and target models (i.e., self-speculation). SparseSpec features a novel sparse attention mechanism, PillarAttn, as the draft model, which accurately selects critical tokens via elegantly reusing information from the verification stage. Furthermore, SparseSpec co-designs self-speculation with three system innovations: (1) a unified scheduler to batch token drafting and verification, (2) delayed verification for CPU/GPU overlap, and (3) dynamic KV-Cache management to maximize memory utilization. Across various models and datasets, SparseSpec outperforms state-of-the-art solutions, with an up to 2.13x throughput speedup.