diff --git a/feed_weekly.xml b/feed_weekly.xml index 30b1afc..2013ce4 100644 --- a/feed_weekly.xml +++ b/feed_weekly.xml @@ -7,741 +7,741 @@ http://www.rssboard.org/rss-specification python-feedgen en - Mon, 01 Dec 2025 00:14:32 +0000 + Mon, 08 Dec 2025 00:12:48 +0000 - VisMem: Latent Vision Memory Unlocks Potential of Vision-Language Models - https://arxiv.org/abs/2511.11007 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.11007.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Xinlei Yu, Chengming Xu, Guibin Zhang, Zhangquan Chen, Yudong Zhang, Yongbo He, Peng-Tao Jiang, Jiangning Zhang, Xiaobin Hu, Shuicheng Yan</p><p><b>Upvotes:</b> 15</p><p><b>Summary:</b> Despite the remarkable success of Vision-Language Models (VLMs), their performance on a range of complex visual tasks is often hindered by a "visual processing bottleneck": a propensity to lose grounding in visual evidence and exhibit a deficit in contextualized visual experience during prolonged generation. Drawing inspiration from human cognitive memory theory, which distinguishes short-term visually-dominant memory and long-term semantically-dominant memory, we propose VisMem, a cognitively-aligned framework that equips VLMs with dynamic latent vision memories, a short-term module for fine-grained perceptual retention and a long-term module for abstract semantic consolidation. These memories are seamlessly invoked during inference, allowing VLMs to maintain both perceptual fidelity and semantic consistency across thinking and generation. Extensive experiments across diverse visual benchmarks for understanding, reasoning, and generation reveal that VisMem delivers a significant average performance boost of 11.8% relative to the vanilla model and outperforms all counterparts, establishing a new paradigm for latent-space memory enhancement. The code will be available: https://github.com/YU-deep/VisMem.git.</p> - https://arxiv.org/abs/2511.11007 - Fri, 14 Nov 2025 06:51:34 +0000 + Where Culture Fades: Revealing the Cultural Gap in Text-to-Image Generation + https://arxiv.org/abs/2511.17282 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.17282.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Chuancheng Shi, Shangze Li, Shiming Guo, Simiao Xie, Wenhua Wu, Jingtong Dou, Chao Wu, Canran Xiao, Cong Wang, Zifeng Cheng, Fei Shen, Tat-Seng Chua</p><p><b>Upvotes:</b> 14</p><p><b>Summary:</b> Multilingual text-to-image (T2I) models have advanced rapidly in terms of visual realism and semantic alignment, and are now widely utilized. Yet outputs vary across cultural contexts: because language carries cultural connotations, images synthesized from multilingual prompts should preserve cross-lingual cultural consistency. We conduct a comprehensive analysis showing that current T2I models often produce culturally neutral or English-biased results under multilingual prompts. Analyses of two representative models indicate that the issue stems not from missing cultural knowledge but from insufficient activation of culture-related representations. We propose a probing method that localizes culture-sensitive signals to a small set of neurons in a few fixed layers. Guided by this finding, we introduce two complementary alignment strategies: (1) inference-time cultural activation that amplifies the identified neurons without backbone fine-tuned; and (2) layer-targeted cultural enhancement that updates only culturally relevant layers. Experiments on our CultureBench demonstrate consistent improvements over strong baselines in cultural consistency while preserving fidelity and diversity.</p> + https://arxiv.org/abs/2511.17282 + Fri, 21 Nov 2025 14:40:50 +0000 - Multi-Agent Deep Research: Training Multi-Agent Systems with M-GRPO - https://arxiv.org/abs/2511.13288 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.13288.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Haoyang Hong, Jiajun Yin, Yuan Wang, Jingnan Liu, Zhe Chen, Ailing Yu, Ji Li, Zhiling Ye, Hansong Xiao, Yefei Chen, Hualei Zhou, Yun Yue, Minghui Yang, Chunxiao Guo, Junwei Liu, Peng Wei, Jinjie Gu</p><p><b>Upvotes:</b> 17</p><p><b>Summary:</b> Multi-agent systems perform well on general reasoning tasks. However, the lack of training in specialized areas hinders their accuracy. Current training methods train a unified large language model (LLM) for all agents in the system. This may limit the performances due to different distributions underlying for different agents. Therefore, training multi-agent systems with distinct LLMs should be the next step to solve. However, this approach introduces optimization challenges. For example, agents operate at different frequencies, rollouts involve varying sub-agent invocations, and agents are often deployed across separate servers, disrupting end-to-end gradient flow. To address these issues, we propose M-GRPO, a hierarchical extension of Group Relative Policy Optimization designed for vertical Multi-agent systems with a main agent (planner) and multiple sub-agents (multi-turn tool executors). M-GRPO computes group-relative advantages for both main and sub-agents, maintaining hierarchical credit assignment. It also introduces a trajectory-alignment scheme that generates fixed-size batches despite variable sub-agent invocations. We deploy a decoupled training pipeline in which agents run on separate servers and exchange minimal statistics via a shared store. This enables scalable training without cross-server backpropagation. In experiments on real-world benchmarks (e.g., GAIA, XBench-DeepSearch, and WebWalkerQA), M-GRPO consistently outperforms both single-agent GRPO and multi-agent GRPO with frozen sub-agents, demonstrating improved stability and sample efficiency. These results show that aligning heterogeneous trajectories and decoupling optimization across specialized agents enhances tool-augmented reasoning tasks.</p> - https://arxiv.org/abs/2511.13288 - Mon, 17 Nov 2025 12:06:30 +0000 + From Code Foundation Models to Agents and Applications: A Practical Guide to Code Intelligence + https://arxiv.org/abs/2511.18538 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.18538.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Jian Yang, Xianglong Liu, Weifeng Lv, Ken Deng, Shawn Guo, Lin Jing, Yizhi Li, Shark Liu, Xianzhen Luo, Yuyu Luo, Changzai Pan, Ensheng Shi, Yingshui Tan, Renshuai Tao, Jiajun Wu, Xianjie Wu, Zhenhe Wu, Daoguang Zan, Chenchen Zhang, Wei Zhang, He Zhu, Terry Yue Zhuo, Kerui Cao, Xianfu Cheng, Jun Dong, Shengjie Fang, Zhiwei Fei, Xiangyuan Guan, Qipeng Guo, Zhiguang Han, Joseph James, Tianqi Luo, Renyuan Li, Yuhang Li, Yiming Liang, Congnan Liu, Jiaheng Liu, Qian Liu, Ruitong Liu, Tyler Loakman, Xiangxin Meng, Chuang Peng, Tianhao Peng, Jiajun Shi, Mingjie Tang, Boyang Wang, Haowen Wang, Yunli Wang, Fanglin Xu, Zihan Xu, Fei Yuan, Ge Zhang, Jiayi Zhang, Xinhao Zhang, Wangchunshu Zhou, Hualei Zhu, King Zhu, Brown Dai, Aishan Liu, Zhoujun Li, Chenghua Lin, Tianyu Liu, Chao Peng, Kai Shen, Libo Qin, Shuangyong Song, Zizheng Zhan, Jiajun Zhang, Jie Zhang, Zhaoxiang Zhang, Bo Zheng</p><p><b>Upvotes:</b> 240</p><p><b>Summary:</b> Large language models (LLMs) have fundamentally transformed automated software development by enabling direct translation of natural language descriptions into functional code, driving commercial adoption through tools like Github Copilot (Microsoft), Cursor (Anysphere), Trae (ByteDance), and Claude Code (Anthropic). While the field has evolved dramatically from rule-based systems to Transformer-based architectures, achieving performance improvements from single-digit to over 95\% success rates on benchmarks like HumanEval. In this work, we provide a comprehensive synthesis and practical guide (a series of analytic and probing experiments) about code LLMs, systematically examining the complete model life cycle from data curation to post-training through advanced prompting paradigms, code pre-training, supervised fine-tuning, reinforcement learning, and autonomous coding agents. We analyze the code capability of the general LLMs (GPT-4, Claude, LLaMA) and code-specialized LLMs (StarCoder, Code LLaMA, DeepSeek-Coder, and QwenCoder), critically examining the techniques, design decisions, and trade-offs. Further, we articulate the research-practice gap between academic research (e.g., benchmarks and tasks) and real-world deployment (e.g., software-related code tasks), including code correctness, security, contextual awareness of large codebases, and integration with development workflows, and map promising research directions to practical needs. Last, we conduct a series of experiments to provide a comprehensive analysis of code pre-training, supervised fine-tuning, and reinforcement learning, covering scaling law, framework selection, hyperparameter sensitivity, model architectures, and dataset comparisons.</p> + https://arxiv.org/abs/2511.18538 + Sun, 23 Nov 2025 17:09:34 +0000 - GigaEvo: An Open Source Optimization Framework Powered By LLMs And Evolution Algorithms - https://arxiv.org/abs/2511.17592 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.17592.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Valentin Khrulkov, Andrey Galichin, Denis Bashkirov, Dmitry Vinichenko, Oleg Travkin, Roman Alferov, Andrey Kuznetsov, Ivan Oseledets</p><p><b>Upvotes:</b> 114</p><p><b>Summary:</b> Recent advances in LLM-guided evolutionary computation, particularly AlphaEvolve (Novikov et al., 2025; Georgiev et al., 2025), have demonstrated remarkable success in discovering novel mathematical constructions and solving challenging optimization problems. However, the high-level descriptions in published work leave many implementation details unspecified, hindering reproducibility and further research. In this report we present GigaEvo, an extensible open-source framework that enables researchers to study and experiment with hybrid LLM-evolution approaches inspired by AlphaEvolve. Our system provides modular implementations of key components: MAP-Elites quality-diversity algorithms, asynchronous DAG-based evaluation pipelines, LLM-driven mutation operators with insight generation and bidirectional lineage tracking, and flexible multi-island evolutionary strategies. In order to assess reproducibility and validate our implementation we evaluate GigaEvo on challenging problems from the AlphaEvolve paper: Heilbronn triangle placement, circle packing in squares, and high-dimensional kissing numbers. The framework emphasizes modularity, concurrency, and ease of experimentation, enabling rapid prototyping through declarative configuration. We provide detailed descriptions of system architecture, implementation decisions, and experimental methodology to support further research in LLM driven evolutionary methods. The GigaEvo framework and all experimental code are available at https://github.com/AIRI-Institute/gigaevo-core.</p> - https://arxiv.org/abs/2511.17592 - Mon, 17 Nov 2025 14:44:47 +0000 + LATTICE: Democratize High-Fidelity 3D Generation at Scale + https://arxiv.org/abs/2512.03052 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2512.03052.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Zeqiang Lai, Yunfei Zhao, Zibo Zhao, Haolin Liu, Qingxiang Lin, Jingwei Huang, Chunchao Guo, Xiangyu Yue</p><p><b>Upvotes:</b> 7</p><p><b>Summary:</b> 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.</p> + https://arxiv.org/abs/2512.03052 + Mon, 24 Nov 2025 03:22:19 +0000 - O-Mem: Omni Memory System for Personalized, Long Horizon, Self-Evolving Agents - https://arxiv.org/abs/2511.13593 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.13593.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Piaohong Wang, Motong Tian, Jiaxian Li, Yuan Liang, Yuqing Wang, Qianben Chen, Tiannan Wang, Zhicong Lu, Jiawei Ma, Yuchen Eleanor Jiang, Wangchunshu Zhou</p><p><b>Upvotes:</b> 23</p><p><b>Summary:</b> Recent advancements in LLM-powered agents have demonstrated significant potential in generating human-like responses; however, they continue to face challenges in maintaining long-term interactions within complex environments, primarily due to limitations in contextual consistency and dynamic personalization. Existing memory systems often depend on semantic grouping prior to retrieval, which can overlook semantically irrelevant yet critical user information and introduce retrieval noise. In this report, we propose the initial design of O-Mem, a novel memory framework based on active user profiling that dynamically extracts and updates user characteristics and event records from their proactive interactions with agents. O-Mem supports hierarchical retrieval of persona attributes and topic-related context, enabling more adaptive and coherent personalized responses. O-Mem achieves 51.67% on the public LoCoMo benchmark, a nearly 3% improvement upon LangMem,the previous state-of-the-art, and it achieves 62.99% on PERSONAMEM, a 3.5% improvement upon A-Mem,the previous state-of-the-art. O-Mem also boosts token and interaction response time efficiency compared to previous memory frameworks. Our work opens up promising directions for developing efficient and human-like personalized AI assistants in the future.</p> - https://arxiv.org/abs/2511.13593 - Mon, 17 Nov 2025 16:55:19 +0000 + DiP: Taming Diffusion Models in Pixel Space + https://arxiv.org/abs/2511.18822 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.18822.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Zhennan Chen, Junwei Zhu, Xu Chen, Jiangning Zhang, Xiaobin Hu, Hanzhen Zhao, Chengjie Wang, Jian Yang, Ying Tai</p><p><b>Upvotes:</b> 25</p><p><b>Summary:</b> Diffusion models face a fundamental trade-off between generation quality and computational efficiency. Latent Diffusion Models (LDMs) offer an efficient solution but suffer from potential information loss and non-end-to-end training. In contrast, existing pixel space models bypass VAEs but are computationally prohibitive for high-resolution synthesis. To resolve this dilemma, we propose DiP, an efficient pixel space diffusion framework. DiP decouples generation into a global and a local stage: a Diffusion Transformer (DiT) backbone operates on large patches for efficient global structure construction, while a co-trained lightweight Patch Detailer Head leverages contextual features to restore fine-grained local details. This synergistic design achieves computational efficiency comparable to LDMs without relying on a VAE. DiP is accomplished with up to 10times faster inference speeds than previous method while increasing the total number of parameters by only 0.3%, and achieves an 1.79 FID score on ImageNet 256times256.</p> + https://arxiv.org/abs/2511.18822 + Mon, 24 Nov 2025 06:55:49 +0000 - MergeDNA: Context-aware Genome Modeling with Dynamic Tokenization through Token Merging - https://arxiv.org/abs/2511.14806 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.14806.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Siyuan Li, Kai Yu, Anna Wang, Zicheng Liu, Chang Yu, Jingbo Zhou, Qirong Yang, Yucheng Guo, Xiaoming Zhang, Stan Z. Li</p><p><b>Upvotes:</b> 8</p><p><b>Summary:</b> Modeling genomic sequences faces two unsolved challenges: the information density varies widely across different regions, while there is no clearly defined minimum vocabulary unit. Relying on either four primitive bases or independently designed DNA tokenizers, existing approaches with naive masked language modeling pre-training often fail to adapt to the varying complexities of genomic sequences. Leveraging Token Merging techniques, this paper introduces a hierarchical architecture that jointly optimizes a dynamic genomic tokenizer and latent Transformers with context-aware pre-training tasks. As for network structures, the tokenization module automatically chunks adjacent bases into words by stacking multiple layers of the differentiable token merging blocks with local-window constraints, then a Latent Encoder captures the global context of these merged words by full-attention blocks. Symmetrically employing a Latent Decoder and a Local Decoder, MergeDNA learns with two pre-training tasks: Merged Token Reconstruction simultaneously trains the dynamic tokenization module and adaptively filters important tokens, while Adaptive Masked Token Modeling learns to predict these filtered tokens to capture informative contents. Extensive experiments show that MergeDNA achieves superior performance on three popular DNA benchmarks and several multi-omics tasks with fine-tuning or zero-shot evaluation, outperforming typical tokenization methods and large-scale DNA foundation models.</p> - https://arxiv.org/abs/2511.14806 - Mon, 17 Nov 2025 19:27:41 +0000 + Nemotron-Flash: Towards Latency-Optimal Hybrid Small Language Models + https://arxiv.org/abs/2511.18890 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.18890.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Yonggan Fu, Xin Dong, Shizhe Diao, Matthijs Van keirsbilck, Hanrong Ye, Wonmin Byeon, Yashaswi Karnati, Lucas Liebenwein, Hannah Zhang, Nikolaus Binder, Maksim Khadkevich, Alexander Keller, Jan Kautz, Yingyan Celine Lin, Pavlo Molchanov</p><p><b>Upvotes:</b> 29</p><p><b>Summary:</b> Efficient deployment of small language models (SLMs) is essential for numerous real-world applications with stringent latency constraints. While previous work on SLM design has primarily focused on reducing the number of parameters to achieve parameter-optimal SLMs, parameter efficiency does not necessarily translate into proportional real-device speed-ups. This work aims to identify the key determinants of SLMs' real-device latency and offer generalizable principles and methodologies for SLM design and training when real-device latency is the primary consideration. Specifically, we identify two central architectural factors: depth-width ratios and operator choices. The former is crucial for small-batch-size latency, while the latter affects both latency and large-batch-size throughput. In light of this, we first study latency-optimal depth-width ratios, with the key finding that although deep-thin models generally achieve better accuracy under the same parameter budget, they may not lie on the accuracy-latency trade-off frontier. Next, we explore emerging efficient attention alternatives to evaluate their potential as candidate building operators. Using the identified promising operators, we construct an evolutionary search framework to automatically discover latency-optimal combinations of these operators within hybrid SLMs, thereby advancing the accuracy-latency frontier. In addition to architectural improvements, we further enhance SLM training using a weight normalization technique that enables more effective weight updates and improves final convergence. Combining these methods, we introduce a new family of hybrid SLMs, called Nemotron-Flash, which significantly advances the accuracy-efficiency frontier of state-of-the-art SLMs, e.g., achieving over +5.5% average accuracy, 1.3x/1.9x lower latency, and 18.7x/45.6x higher throughput compared to Qwen3-1.7B/0.6B, respectively.</p> + https://arxiv.org/abs/2511.18890 + Mon, 24 Nov 2025 08:46:36 +0000 - InstructMix2Mix: Consistent Sparse-View Editing Through Multi-View Model Personalization - https://arxiv.org/abs/2511.14899 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.14899.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Daniel Gilo, Or Litany</p><p><b>Upvotes:</b> 11</p><p><b>Summary:</b> We address the task of multi-view image editing from sparse input views, where the inputs can be seen as a mix of images capturing the scene from different viewpoints. The goal is to modify the scene according to a textual instruction while preserving consistency across all views. Existing methods, based on per-scene neural fields or temporal attention mechanisms, struggle in this setting, often producing artifacts and incoherent edits. We propose InstructMix2Mix (I-Mix2Mix), a framework that distills the editing capabilities of a 2D diffusion model into a pretrained multi-view diffusion model, leveraging its data-driven 3D prior for cross-view consistency. A key contribution is replacing the conventional neural field consolidator in Score Distillation Sampling (SDS) with a multi-view diffusion student, which requires novel adaptations: incremental student updates across timesteps, a specialized teacher noise scheduler to prevent degeneration, and an attention modification that enhances cross-view coherence without additional cost. Experiments demonstrate that I-Mix2Mix significantly improves multi-view consistency while maintaining high per-frame edit quality.</p> - https://arxiv.org/abs/2511.14899 - Tue, 18 Nov 2025 20:37:52 +0000 + Deep Research: A Systematic Survey + https://arxiv.org/abs/2512.02038 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2512.02038.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Zhengliang Shi, Yiqun Chen, Haitao Li, Weiwei Sun, Shiyu Ni, Yougang Lyu, Run-Ze Fan, Bowen Jin, Yixuan Weng, Minjun Zhu, Qiujie Xie, Xinyu Guo, Qu Yang, Jiayi Wu, Jujia Zhao, Xiaqiang Tang, Xinbei Ma, Cunxiang Wang, Jiaxin Mao, Qingyao Ai, Jen-Tse Huang, Wenxuan Wang, Yue Zhang, Yiming Yang, Zhaopeng Tu, Zhaochun Ren</p><p><b>Upvotes:</b> 55</p><p><b>Summary:</b> 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.</p> + https://arxiv.org/abs/2512.02038 + Mon, 24 Nov 2025 15:28:28 +0000 - Unveiling Intrinsic Dimension of Texts: from Academic Abstract to Creative Story - https://arxiv.org/abs/2511.15210 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.15210.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Vladislav Pedashenko, Laida Kushnareva, Yana Khassan Nibal, Eduard Tulchinskii, Kristian Kuznetsov, Vladislav Zharchinskii, Yury Maximov, Irina Piontkovskaya</p><p><b>Upvotes:</b> 86</p><p><b>Summary:</b> Intrinsic dimension (ID) is an important tool in modern LLM analysis, informing studies of training dynamics, scaling behavior, and dataset structure, yet its textual determinants remain underexplored. We provide the first comprehensive study grounding ID in interpretable text properties through cross-encoder analysis, linguistic features, and sparse autoencoders (SAEs). In this work, we establish three key findings. First, ID is complementary to entropy-based metrics: after controlling for length, the two are uncorrelated, with ID capturing geometric complexity orthogonal to prediction quality. Second, ID exhibits robust genre stratification: scientific prose shows low ID (~8), encyclopedic content medium ID (~9), and creative/opinion writing high ID (~10.5) across all models tested. This reveals that contemporary LLMs find scientific text "representationally simple" while fiction requires additional degrees of freedom. Third, using SAEs, we identify causal features: scientific signals (formal tone, report templates, statistics) reduce ID; humanized signals (personalization, emotion, narrative) increase it. Steering experiments confirm these effects are causal. Thus, for contemporary models, scientific writing appears comparatively "easy", whereas fiction, opinion, and affect add representational degrees of freedom. Our multi-faceted analysis provides practical guidance for the proper use of ID and the sound interpretation of ID-based results.</p> - https://arxiv.org/abs/2511.15210 - Wed, 19 Nov 2025 08:00:40 +0000 + Mixture of Horizons in Action Chunking + https://arxiv.org/abs/2511.19433 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.19433.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Dong Jing, Gang Wang, Jiaqi Liu, Weiliang Tang, Zelong Sun, Yunchao Yao, Zhenyu Wei, Yunhui Liu, Zhiwu Lu, Mingyu Ding</p><p><b>Upvotes:</b> 17</p><p><b>Summary:</b> Vision-language-action (VLA) models have shown remarkable capabilities in robotic manipulation, but their performance is sensitive to the action chunk length used during training, termed horizon. Our empirical study reveals an inherent trade-off: longer horizons provide stronger global foresight but degrade fine-grained accuracy, while shorter ones sharpen local control yet struggle on long-term tasks, implying fixed choice of single horizons being suboptimal. To mitigate the trade-off, we propose a mixture of horizons (MoH) strategy. MoH rearranges the action chunk into several segments with different horizons, processes them in parallel with a shared action transformer, and fuses outputs with a light linear gate. It has three appealing benefits. 1) MoH exploits long-term foresight and short-term precision jointly within a single model, improving both performance and generalizability to complex tasks. 2) MoH is plug-and-play for full-attention action modules with minimal training or inference overhead. 3) MoH enables dynamic inference with adaptive horizons, which selects stable actions through cross-horizon consensus, achieving 2.5times higher throughput than baselines while preserving superior performance. Extensive experiments over flow-based policies π_0, π_{0.5}, and one-step regression policy π_{reg} demonstrate that MoH yields consistent and significant gains on both simulations and real-world tasks. Notably, under mixed-task setting, π_{0.5} with MoH reaches a new state-of-the-art with 99% average success rate on LIBERO after only 30k training iterations. Project page: https://github.com/Timsty1/MixtureOfHorizons</p> + https://arxiv.org/abs/2511.19433 + Mon, 24 Nov 2025 18:59:51 +0000 - Taming Generative Synthetic Data for X-ray Prohibited Item Detection - https://arxiv.org/abs/2511.15299 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.15299.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Jialong Sun, Hongguang Zhu, Weizhe Liu, Yunda Sun, Renshuai Tao, Yunchao Wei</p><p><b>Upvotes:</b> 2</p><p><b>Summary:</b> Training prohibited item detection models requires a large amount of X-ray security images, but collecting and annotating these images is time-consuming and laborious. To address data insufficiency, X-ray security image synthesis methods composite images to scale up datasets. However, previous methods primarily follow a two-stage pipeline, where they implement labor-intensive foreground extraction in the first stage and then composite images in the second stage. Such a pipeline introduces inevitable extra labor cost and is not efficient. In this paper, we propose a one-stage X-ray security image synthesis pipeline (Xsyn) based on text-to-image generation, which incorporates two effective strategies to improve the usability of synthetic images. The Cross-Attention Refinement (CAR) strategy leverages the cross-attention map from the diffusion model to refine the bounding box annotation. The Background Occlusion Modeling (BOM) strategy explicitly models background occlusion in the latent space to enhance imaging complexity. To the best of our knowledge, compared with previous methods, Xsyn is the first to achieve high-quality X-ray security image synthesis without extra labor cost. Experiments demonstrate that our method outperforms all previous methods with 1.2% mAP improvement, and the synthetic images generated by our method are beneficial to improve prohibited item detection performance across various X-ray security datasets and detectors. Code is available at https://github.com/pILLOW-1/Xsyn/.</p> - https://arxiv.org/abs/2511.15299 - Wed, 19 Nov 2025 10:07:11 +0000 + The Curious Case of Analogies: Investigating Analogical Reasoning in Large Language Models + https://arxiv.org/abs/2511.20344 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.20344.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Taewhoo Lee, Minju Song, Chanwoong Yoon, Jungwoo Park, Jaewoo Kang</p><p><b>Upvotes:</b> 11</p><p><b>Summary:</b> Analogical reasoning is at the core of human cognition, serving as an important foundation for a variety of intellectual activities. While prior work has shown that LLMs can represent task patterns and surface-level concepts, it remains unclear whether these models can encode high-level relational concepts and apply them to novel situations through structured comparisons. In this work, we explore this fundamental aspect using proportional and story analogies, and identify three key findings. First, LLMs effectively encode the underlying relationships between analogous entities; both attributive and relational information propagate through mid-upper layers in correct cases, whereas reasoning failures reflect missing relational information within these layers. Second, unlike humans, LLMs often struggle not only when relational information is missing, but also when attempting to apply it to new entities. In such cases, strategically patching hidden representations at critical token positions can facilitate information transfer to a certain extent. Lastly, successful analogical reasoning in LLMs is marked by strong structural alignment between analogous situations, whereas failures often reflect degraded or misplaced alignment. Overall, our findings reveal that LLMs exhibit emerging but limited capabilities in encoding and applying high-level relational concepts, highlighting both parallels and gaps with human cognition.</p> + https://arxiv.org/abs/2511.20344 + Tue, 25 Nov 2025 14:23:58 +0000 - Insights from the ICLR Peer Review and Rebuttal Process - https://arxiv.org/abs/2511.15462 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.15462.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Amir Hossein Kargaran, Nafiseh Nikeghbal, Jing Yang, Nedjma Ousidhoum</p><p><b>Upvotes:</b> 6</p><p><b>Summary:</b> Peer review is a cornerstone of scientific publishing, including at premier machine learning conferences such as ICLR. As submission volumes increase, understanding the nature and dynamics of the review process is crucial for improving its efficiency, effectiveness, and the quality of published papers. We present a large-scale analysis of the ICLR 2024 and 2025 peer review processes, focusing on before- and after-rebuttal scores and reviewer-author interactions. We examine review scores, author-reviewer engagement, temporal patterns in review submissions, and co-reviewer influence effects. Combining quantitative analyses with LLM-based categorization of review texts and rebuttal discussions, we identify common strengths and weaknesses for each rating group, as well as trends in rebuttal strategies that are most strongly associated with score changes. Our findings show that initial scores and the ratings of co-reviewers are the strongest predictors of score changes during the rebuttal, pointing to a degree of reviewer influence. Rebuttals play a valuable role in improving outcomes for borderline papers, where thoughtful author responses can meaningfully shift reviewer perspectives. More broadly, our study offers evidence-based insights to improve the peer review process, guiding authors on effective rebuttal strategies and helping the community design fairer and more efficient review processes. Our code and score changes data are available at https://github.com/papercopilot/iclr-insights.</p> - https://arxiv.org/abs/2511.15462 - Wed, 19 Nov 2025 14:21:52 +0000 + Flash-DMD: Towards High-Fidelity Few-Step Image Generation with Efficient Distillation and Joint Reinforcement Learning + https://arxiv.org/abs/2511.20549 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.20549.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Guanjie Chen, Shirui Huang, Kai Liu, Jianchen Zhu, Xiaoye Qu, Peng Chen, Yu Cheng, Yifu Sun</p><p><b>Upvotes:</b> 23</p><p><b>Summary:</b> Diffusion Models have emerged as a leading class of generative models, yet their iterative sampling process remains computationally expensive. Timestep distillation is a promising technique to accelerate generation, but it often requires extensive training and leads to image quality degradation. Furthermore, fine-tuning these distilled models for specific objectives, such as aesthetic appeal or user preference, using Reinforcement Learning (RL) is notoriously unstable and easily falls into reward hacking. In this work, we introduce Flash-DMD, a novel framework that enables fast convergence with distillation and joint RL-based refinement. Specifically, we first propose an efficient timestep-aware distillation strategy that significantly reduces training cost with enhanced realism, outperforming DMD2 with only 2.1% its training cost. Second, we introduce a joint training scheme where the model is fine-tuned with an RL objective while the timestep distillation training continues simultaneously. We demonstrate that the stable, well-defined loss from the ongoing distillation acts as a powerful regularizer, effectively stabilizing the RL training process and preventing policy collapse. Extensive experiments on score-based and flow matching models show that our proposed Flash-DMD not only converges significantly faster but also achieves state-of-the-art generation quality in the few-step sampling regime, outperforming existing methods in visual quality, human preference, and text-image alignment metrics. Our work presents an effective paradigm for training efficient, high-fidelity, and stable generative models. Codes are coming soon.</p> + https://arxiv.org/abs/2511.20549 + Tue, 25 Nov 2025 17:47:11 +0000 - Multimodal Evaluation of Russian-language Architectures - https://arxiv.org/abs/2511.15552 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.15552.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Artem Chervyakov, Ulyana Isaeva, Anton Emelyanov, Artem Safin, Maria Tikhonova, Alexander Kharitonov, Yulia Lyakh, Petr Surovtsev, Denis Shevelev, Vildan Saburov, Vasily Konovalov, Elisei Rykov, Ivan Sviridov, Amina Miftakhova, Ilseyar Alimova, Alexander Panchenko, Alexander Kapitanov, Alena Fenogenova</p><p><b>Upvotes:</b> 74</p><p><b>Summary:</b> Multimodal large language models (MLLMs) are currently at the center of research attention, showing rapid progress in scale and capabilities, yet their intelligence, limitations, and risks remain insufficiently understood. To address these issues, particularly in the context of the Russian language, where no multimodal benchmarks currently exist, we introduce Mera Multi, an open multimodal evaluation framework for Russian-spoken architectures. The benchmark is instruction-based and encompasses default text, image, audio, and video modalities, comprising 18 newly constructed evaluation tasks for both general-purpose models and modality-specific architectures (image-to-text, video-to-text, and audio-to-text). Our contributions include: (i) a universal taxonomy of multimodal abilities; (ii) 18 datasets created entirely from scratch with attention to Russian cultural and linguistic specificity, unified prompts, and metrics; (iii) baseline results for both closed-source and open-source models; (iv) a methodology for preventing benchmark leakage, including watermarking and licenses for private sets. While our current focus is on Russian, the proposed benchmark provides a replicable methodology for constructing multimodal benchmarks in typologically diverse languages, particularly within the Slavic language family.</p> - https://arxiv.org/abs/2511.15552 - Wed, 19 Nov 2025 15:43:53 +0000 + The Consistency Critic: Correcting Inconsistencies in Generated Images via Reference-Guided Attentive Alignment + https://arxiv.org/abs/2511.20614 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.20614.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Ziheng Ouyang, Yiren Song, Yaoli Liu, Shihao Zhu, Qibin Hou, Ming-Ming Cheng, Mike Zheng Shou</p><p><b>Upvotes:</b> 37</p><p><b>Summary:</b> Previous works have explored various customized generation tasks given a reference image, but they still face limitations in generating consistent fine-grained details. In this paper, our aim is to solve the inconsistency problem of generated images by applying a reference-guided post-editing approach and present our ImageCritic. We first construct a dataset of reference-degraded-target triplets obtained via VLM-based selection and explicit degradation, which effectively simulates the common inaccuracies or inconsistencies observed in existing generation models. Furthermore, building on a thorough examination of the model's attention mechanisms and intrinsic representations, we accordingly devise an attention alignment loss and a detail encoder to precisely rectify inconsistencies. ImageCritic can be integrated into an agent framework to automatically detect inconsistencies and correct them with multi-round and local editing in complex scenarios. Extensive experiments demonstrate that ImageCritic can effectively resolve detail-related issues in various customized generation scenarios, providing significant improvements over existing methods.</p> + https://arxiv.org/abs/2511.20614 + Tue, 25 Nov 2025 18:40:25 +0000 - Computer-Use Agents as Judges for Generative User Interface - https://arxiv.org/abs/2511.15567 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.15567.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Kevin Qinghong Lin, Siyuan Hu, Linjie Li, Zhengyuan Yang, Lijuan Wang, Philip Torr, Mike Zheng Shou</p><p><b>Upvotes:</b> 50</p><p><b>Summary:</b> Computer-Use Agents (CUA) are becoming increasingly capable of autonomously operating digital environments through Graphical User Interfaces (GUI). Yet, most GUI remain designed primarily for humans--prioritizing aesthetics and usability--forcing agents to adopt human-oriented behaviors that are unnecessary for efficient task execution. At the same time, rapid advances in coding-oriented language models (Coder) have transformed automatic GUI design. This raises a fundamental question: Can CUA as judges to assist Coder for automatic GUI design? To investigate, we introduce AUI-Gym, a benchmark for Automatic GUI development spanning 52 applications across diverse domains. Using language models, we synthesize 1560 tasks that simulate real-world scenarios. To ensure task reliability, we further develop a verifier that programmatically checks whether each task is executable within its environment. Building on this, we propose a Coder-CUA in Collaboration framework: the Coder acts as Designer, generating and revising websites, while the CUA serves as Judge, evaluating functionality and refining designs. Success is measured not by visual appearance, but by task solvability and CUA navigation success rate. To turn CUA feedback into usable guidance, we design a CUA Dashboard that compresses multi-step navigation histories into concise visual summaries, offering interpretable guidance for iterative redesign. By positioning agents as both designers and judges, our framework shifts interface design toward agent-native efficiency and reliability. Our work takes a step toward shifting agents from passive use toward active participation in digital environments. Our code and dataset are available at https://github.com/showlab/AUI.</p> - https://arxiv.org/abs/2511.15567 - Wed, 19 Nov 2025 16:00:02 +0000 + PixelDiT: Pixel Diffusion Transformers for Image Generation + https://arxiv.org/abs/2511.20645 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.20645.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Yongsheng Yu, Wei Xiong, Weili Nie, Yichen Sheng, Shiqiu Liu, Jiebo Luo</p><p><b>Upvotes:</b> 25</p><p><b>Summary:</b> Latent-space modeling has been the standard for Diffusion Transformers (DiTs). However, it relies on a two-stage pipeline where the pretrained autoencoder introduces lossy reconstruction, leading to error accumulation while hindering joint optimization. To address these issues, we propose PixelDiT, a single-stage, end-to-end model that eliminates the need for the autoencoder and learns the diffusion process directly in the pixel space. PixelDiT adopts a fully transformer-based architecture shaped by a dual-level design: a patch-level DiT that captures global semantics and a pixel-level DiT that refines texture details, enabling efficient training of a pixel-space diffusion model while preserving fine details. Our analysis reveals that effective pixel-level token modeling is essential to the success of pixel diffusion. PixelDiT achieves 1.61 FID on ImageNet 256x256, surpassing existing pixel generative models by a large margin. We further extend PixelDiT to text-to-image generation and pretrain it at the 1024x1024 resolution in pixel space. It achieves 0.74 on GenEval and 83.5 on DPG-bench, approaching the best latent diffusion models.</p> + https://arxiv.org/abs/2511.20645 + Tue, 25 Nov 2025 18:59:25 +0000 - Think Visually, Reason Textually: Vision-Language Synergy in ARC - https://arxiv.org/abs/2511.15703 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.15703.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Beichen Zhang, Yuhang Zang, Xiaoyi Dong, Yuhang Cao, Haodong Duan, Dahua Lin, Jiaqi Wang</p><p><b>Upvotes:</b> 8</p><p><b>Summary:</b> Abstract reasoning from minimal examples remains a core unsolved problem for frontier foundation models such as GPT-5 and Grok 4. These models still fail to infer structured transformation rules from a handful of examples, which is a key hallmark of human intelligence. The Abstraction and Reasoning Corpus for Artificial General Intelligence (ARC-AGI) provides a rigorous testbed for this capability, demanding conceptual rule induction and transfer to novel tasks. Most existing methods treat ARC-AGI as a purely textual reasoning task, overlooking the fact that humans rely heavily on visual abstraction when solving such puzzles. However, our pilot experiments reveal a paradox: naively rendering ARC-AGI grids as images degrades performance due to imprecise rule execution. This leads to our central hypothesis that vision and language possess complementary strengths across distinct reasoning stages: vision supports global pattern abstraction and verification, whereas language specializes in symbolic rule formulation and precise execution. Building on this insight, we introduce two synergistic strategies: (1) Vision-Language Synergy Reasoning (VLSR), which decomposes ARC-AGI into modality-aligned subtasks; and (2) Modality-Switch Self-Correction (MSSC), which leverages vision to verify text-based reasoning for intrinsic error correction. Extensive experiments demonstrate that our approach yields up to a 4.33% improvement over text-only baselines across diverse flagship models and multiple ARC-AGI tasks. Our findings suggest that unifying visual abstraction with linguistic reasoning is a crucial step toward achieving generalizable, human-like intelligence in future foundation models. Source code will be released soon.</p> - https://arxiv.org/abs/2511.15703 - Wed, 19 Nov 2025 18:59:04 +0000 + Infinity-RoPE: Action-Controllable Infinite Video Generation Emerges From Autoregressive Self-Rollout + https://arxiv.org/abs/2511.20649 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.20649.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Hidir Yesiltepe, Tuna Han Salih Meral, Adil Kaan Akan, Kaan Oktay, Pinar Yanardag</p><p><b>Upvotes:</b> 43</p><p><b>Summary:</b> Current autoregressive video diffusion models are constrained by three core bottlenecks: (i) the finite temporal horizon imposed by the base model's 3D Rotary Positional Embedding (3D-RoPE), (ii) slow prompt responsiveness in maintaining fine-grained action control during long-form rollouts, and (iii) the inability to realize discontinuous cinematic transitions within a single generation stream. We introduce infty-RoPE, a unified inference-time framework that addresses all three limitations through three interconnected components: Block-Relativistic RoPE, KV Flush, and RoPE Cut. Block-Relativistic RoPE reformulates temporal encoding as a moving local reference frame, where each newly generated latent block is rotated relative to the base model's maximum frame horizon while earlier blocks are rotated backward to preserve relative temporal geometry. This relativistic formulation eliminates fixed temporal positions, enabling continuous video generation far beyond the base positional limits. To obtain fine-grained action control without re-encoding, KV Flush renews the KV cache by retaining only two latent frames, the global sink and the last generated latent frame, thereby ensuring immediate prompt responsiveness. Finally, RoPE Cut introduces controlled discontinuities in temporal RoPE coordinates, enabling multi-cut scene transitions within a single continuous rollout. Together, these components establish infty-RoPE as a training-free foundation for infinite-horizon, controllable, and cinematic video diffusion. Comprehensive experiments show that infty-RoPE consistently surpasses previous autoregressive models in overall VBench scores.</p> + https://arxiv.org/abs/2511.20649 + Tue, 25 Nov 2025 18:59:46 +0000 - GeoVista: Web-Augmented Agentic Visual Reasoning for Geolocalization - https://arxiv.org/abs/2511.15705 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.15705.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Yikun Wang, Zuyan Liu, Ziyi Wang, Pengfei Liu, Han Hu, Yongming Rao</p><p><b>Upvotes:</b> 88</p><p><b>Summary:</b> Current research on agentic visual reasoning enables deep multimodal understanding but primarily focuses on image manipulation tools, leaving a gap toward more general-purpose agentic models. In this work, we revisit the geolocalization task, which requires not only nuanced visual grounding but also web search to confirm or refine hypotheses during reasoning. Since existing geolocalization benchmarks fail to meet the need for high-resolution imagery and the localization challenge for deep agentic reasoning, we curate GeoBench, a benchmark that includes photos and panoramas from around the world, along with a subset of satellite images of different cities to rigorously evaluate the geolocalization ability of agentic models. We also propose GeoVista, an agentic model that seamlessly integrates tool invocation within the reasoning loop, including an image-zoom-in tool to magnify regions of interest and a web-search tool to retrieve related web information. We develop a complete training pipeline for it, including a cold-start supervised fine-tuning (SFT) stage to learn reasoning patterns and tool-use priors, followed by a reinforcement learning (RL) stage to further enhance reasoning ability. We adopt a hierarchical reward to leverage multi-level geographical information and improve overall geolocalization performance. Experimental results show that GeoVista surpasses other open-source agentic models on the geolocalization task greatly and achieves performance comparable to closed-source models such as Gemini-2.5-flash and GPT-5 on most metrics.</p> - https://arxiv.org/abs/2511.15705 - Wed, 19 Nov 2025 18:59:22 +0000 + LongVT: Incentivizing "Thinking with Long Videos" via Native Tool Calling + https://arxiv.org/abs/2511.20785 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.20785.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Zuhao Yang, Sudong Wang, Kaichen Zhang, Keming Wu, Sicong Leng, Yifan Zhang, Chengwei Qin, Shijian Lu, Xingxuan Li, Lidong Bing</p><p><b>Upvotes:</b> 148</p><p><b>Summary:</b> Large multimodal models (LMMs) have shown great potential for video reasoning with textual Chain-of-Thought. However, they remain vulnerable to hallucinations, especially when processing long-form videos where evidence is sparse and temporally dispersed. Inspired by how humans comprehend long videos - by first skimming globally and then examining relevant clips for details - we introduce LongVT, an end-to-end agentic framework that enables "Thinking with Long Videos" via interleaved Multimodal Chain-of-Tool-Thought. Specifically, we exploit LMMs' inherent temporal grounding ability as a native video cropping tool to zoom in on a specific video clip and resample finer-grained video frames. This global-to-local reasoning loop continues until answers are grounded in retrieved visual evidence. Given the scarcity of fine-grained question-answering (QA) data for the long video reasoning task, we curate and will release a data suite named VideoSIAH to facilitate both training and evaluation. Specifically, our training dataset consists of 247.9K samples for tool-integrated cold-start supervised fine-tuning, 1.6K samples for agentic reinforcement learning, and 15.4K samples for agentic reinforcement fine-tuning, respectively. Our evaluation benchmark consists of 1,280 QA pairs that are carefully curated through a semi-automatic data pipeline with human-in-the-loop validation. With a meticulously designed three-stage training strategy and extensive empirical validation, LongVT consistently outperforms existing strong baselines across four challenging long-video understanding and reasoning benchmarks. Our codes, data, and model checkpoints are publicly available at https://github.com/EvolvingLMMs-Lab/LongVT .</p> + https://arxiv.org/abs/2511.20785 + Tue, 25 Nov 2025 19:22:48 +0000 - Unified all-atom molecule generation with neural fields - https://arxiv.org/abs/2511.15906 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.15906.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Matthieu Kirchmeyer, Pedro O. Pinheiro, Emma Willett, Karolis Martinkus, Joseph Kleinhenz, Emily K. Makowski, Andrew M. Watkins, Vladimir Gligorijevic, Richard Bonneau, Saeed Saremi</p><p><b>Upvotes:</b> 2</p><p><b>Summary:</b> Generative models for structure-based drug design are often limited to a specific modality, restricting their broader applicability. To address this challenge, we introduce FuncBind, a framework based on computer vision to generate target-conditioned, all-atom molecules across atomic systems. FuncBind uses neural fields to represent molecules as continuous atomic densities and employs score-based generative models with modern architectures adapted from the computer vision literature. This modality-agnostic representation allows a single unified model to be trained on diverse atomic systems, from small to large molecules, and handle variable atom/residue counts, including non-canonical amino acids. FuncBind achieves competitive in silico performance in generating small molecules, macrocyclic peptides, and antibody complementarity-determining region loops, conditioned on target structures. FuncBind also generated in vitro novel antibody binders via de novo redesign of the complementarity-determining region H3 loop of two chosen co-crystal structures. As a final contribution, we introduce a new dataset and benchmark for structure-conditioned macrocyclic peptide generation. The code is available at https://github.com/prescient-design/funcbind.</p> - https://arxiv.org/abs/2511.15906 - Wed, 19 Nov 2025 22:18:13 +0000 + CaptionQA: Is Your Caption as Useful as the Image Itself? + https://arxiv.org/abs/2511.21025 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.21025.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Shijia Yang, Yunong Liu, Bohan Zhai, Ximeng Sun, Zicheng Liu, Emad Barsoum, Manling Li, Chenfeng Xu</p><p><b>Upvotes:</b> 25</p><p><b>Summary:</b> Image captions serve as efficient surrogates for visual content in multimodal systems such as retrieval, recommendation, and multi-step agentic inference pipelines. Yet current evaluation practices miss a fundamental question: Can captions stand-in for images in real downstream tasks? We propose a utility-based benchmark, CaptionQA, to evaluate model-generated captions, where caption quality is measured by how well it supports downstream tasks. CaptionQA is an extensible domain-dependent benchmark covering 4 domains--Natural, Document, E-commerce, and Embodied AI--each with fine-grained taxonomies (25 top-level and 69 subcategories) that identify useful information for domain-specific tasks. CaptionQA builds 33,027 densely annotated multiple-choice questions (50.3 per image on average) that explicitly require visual information to answer, providing a comprehensive probe of caption utility. In our evaluation protocol, an LLM answers these questions using captions alone, directly measuring whether captions preserve image-level utility and are utilizable by a downstream LLM. Evaluating state-of-the-art MLLMs reveals substantial gaps between the image and its caption utility. Notably, models nearly identical on traditional image-QA benchmarks lower by up to 32% in caption utility. We release CaptionQA along with an open-source pipeline for extension to new domains. The code is available at https://github.com/bronyayang/CaptionQA.</p> + https://arxiv.org/abs/2511.21025 + Wed, 26 Nov 2025 03:43:32 +0000 - EvoVLA: Self-Evolving Vision-Language-Action Model - https://arxiv.org/abs/2511.16166 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.16166.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Zeting Liu, Zida Yang, Zeyu Zhang, Hao Tang</p><p><b>Upvotes:</b> 4</p><p><b>Summary:</b> Long-horizon robotic manipulation remains challenging for Vision-Language-Action (VLA) models despite recent progress in zero-shot generalization and simulation-to-real-world transfer. Current VLA models suffer from stage hallucination, where agents exploit coarse evaluation signals to shortcut multi-step tasks, reporting high progress without truly completing them. We present EvoVLA, a self-supervised VLA framework that addresses this issue through three complementary components: Stage-Aligned Reward (SAR), which uses triplet contrastive learning with Gemini-generated hard negatives to prevent visual shortcuts; Pose-Based Object Exploration (POE), which grounds curiosity in relative object-gripper pose instead of raw pixels; and Long-Horizon Memory, which uses selective context retention and gated fusion to stabilize intrinsic shaping during extended rollouts. Extensive evaluations on Discoverse-L, a long-horizon manipulation benchmark with three multi-stage tasks, show that EvoVLA improves average task success by 10.2 percentage points over the strongest baseline (OpenVLA-OFT), reaching 69.2 percent. EvoVLA also achieves one-and-a-half times better sample efficiency and reduces stage hallucination from 38.5 percent to 14.8 percent. Real-world deployment on physical robots reaches an average success rate of 54.6 percent across four manipulation tasks, outperforming OpenVLA-OFT by 11 points, demonstrating effective sim-to-real transfer and strong generalization. Code: https://github.com/AIGeeksGroup/EvoVLA. Website: https://aigeeksgroup.github.io/EvoVLA.</p> - https://arxiv.org/abs/2511.16166 - Thu, 20 Nov 2025 09:08:33 +0000 + Qwen3-VL Technical Report + https://arxiv.org/abs/2511.21631 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.21631.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Shuai Bai, Yuxuan Cai, Ruizhe Chen, Keqin Chen, Xionghui Chen, Zesen Cheng, Lianghao Deng, Wei Ding, Chang Gao, Chunjiang Ge, Wenbin Ge, Zhifang Guo, Qidong Huang, Jie Huang, Fei Huang, Binyuan Hui, Shutong Jiang, Zhaohai Li, Mingsheng Li, Mei Li, Kaixin Li, Zicheng Lin, Junyang Lin, Xuejing Liu, Jiawei Liu, Chenglong Liu, Yang Liu, Dayiheng Liu, Shixuan Liu, Dunjie Lu, Ruilin Luo, Chenxu Lv, Rui Men, Lingchen Meng, Xuancheng Ren, Xingzhang Ren, Sibo Song, Yuchong Sun, Jun Tang, Jianhong Tu, Jianqiang Wan, Peng Wang, Pengfei Wang, Qiuyue Wang, Yuxuan Wang, Tianbao Xie, Yiheng Xu, Haiyang Xu, Jin Xu, Zhibo Yang, Mingkun Yang, Jianxin Yang, An Yang, Bowen Yu, Fei Zhang, Hang Zhang, Xi Zhang, Bo Zheng, Humen Zhong, Jingren Zhou, Fan Zhou, Jing Zhou, Yuanzhi Zhu, Ke Zhu</p><p><b>Upvotes:</b> 107</p><p><b>Summary:</b> We introduce Qwen3-VL, the most capable vision-language model in the Qwen series to date, achieving superior performance across a broad range of multimodal benchmarks. It natively supports interleaved contexts of up to 256K tokens, seamlessly integrating text, images, and video. The model family includes both dense (2B/4B/8B/32B) and mixture-of-experts (30B-A3B/235B-A22B) variants to accommodate diverse latency-quality trade-offs. Qwen3-VL delivers three core pillars: (i) markedly stronger pure-text understanding, surpassing comparable text-only backbones in several cases; (ii) robust long-context comprehension with a native 256K-token window for both text and interleaved multimodal inputs, enabling faithful retention, retrieval, and cross-referencing across long documents and videos; and (iii) advanced multimodal reasoning across single-image, multi-image, and video tasks, demonstrating leading performance on comprehensive evaluations such as MMMU and visual-math benchmarks (e.g., MathVista and MathVision). Architecturally, we introduce three key upgrades: (i) an enhanced interleaved-MRoPE for stronger spatial-temporal modeling across images and video; (ii) DeepStack integration, which effectively leverages multi-level ViT features to tighten vision-language alignment; and (iii) text-based time alignment for video, evolving from T-RoPE to explicit textual timestamp alignment for more precise temporal grounding. Under comparable token budgets and latency constraints, Qwen3-VL achieves superior performance in both dense and Mixture-of-Experts (MoE) architectures. We envision Qwen3-VL serving as a foundational engine for image-grounded reasoning, agentic decision-making, and multimodal code intelligence in real-world workflows.</p> + https://arxiv.org/abs/2511.21631 + Wed, 26 Nov 2025 17:59:08 +0000 - Mantis: A Versatile Vision-Language-Action Model with Disentangled Visual Foresight - https://arxiv.org/abs/2511.16175 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.16175.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Yi Yang, Xueqi Li, Yiyang Chen, Jin Song, Yihan Wang, Zipeng Xiao, Jiadi Su, You Qiaoben, Pengfei Liu, Zhijie Deng</p><p><b>Upvotes:</b> 12</p><p><b>Summary:</b> Recent advances in Vision-Language-Action (VLA) models demonstrate that visual signals can effectively complement sparse action supervisions. However, letting VLA directly predict high-dimensional visual states can distribute model capacity and incur prohibitive training cost, while compressing visual states into more compact supervisory signals inevitably incurs information bottlenecks. Moreover, existing methods often suffer from poor comprehension and reasoning capabilities due to the neglect of language supervision. This paper introduces Mantis, a novel framework featuring a Disentangled Visual Foresight (DVF) to tackle these issues. Specifically, Mantis decouples visual foresight prediction from the backbone with the combination of meta queries and a diffusion Transformer (DiT) head. With the current visual state provided to the DiT via a residual connection, a simple next-state prediction objective enables the meta queries to automatically capture the latent actions that delineate the visual trajectory, and hence boost the learning of explicit actions. The disentanglement reduces the burden of the VLA backbone, enabling it to maintain comprehension and reasoning capabilities through language supervision. Empirically, pretrained on human manipulation videos, robot demonstrations, and image-text pairs, Mantis achieves a 96.7% success rate on LIBERO benchmark after fine-tuning, surpassing powerful baselines while exhibiting high convergence speed. Real-world evaluations show that Mantis outperforms π_{0.5}, a leading open-source VLA model, particularly in instruction-following capability, generalization to unseen instructions, and reasoning ability. Code and weights are released to support the open-source community.</p> - https://arxiv.org/abs/2511.16175 - Thu, 20 Nov 2025 09:30:23 +0000 + ToolOrchestra: Elevating Intelligence via Efficient Model and Tool Orchestration + https://arxiv.org/abs/2511.21689 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.21689.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Hongjin Su, Shizhe Diao, Ximing Lu, Mingjie Liu, Jiacheng Xu, Xin Dong, Yonggan Fu, Peter Belcak, Hanrong Ye, Hongxu Yin, Yi Dong, Evelina Bakhturina, Tao Yu, Yejin Choi, Jan Kautz, Pavlo Molchanov</p><p><b>Upvotes:</b> 96</p><p><b>Summary:</b> Large language models are powerful generalists, yet solving deep and complex problems such as those of the Humanity's Last Exam (HLE) remains both conceptually challenging and computationally expensive. We show that small orchestrators managing other models and a variety of tools can both push the upper bound of intelligence and improve efficiency in solving difficult agentic tasks. We introduce ToolOrchestra, a method for training small orchestrators that coordinate intelligent tools. ToolOrchestra explicitly uses reinforcement learning with outcome-, efficiency-, and user-preference-aware rewards. Using ToolOrchestra, we produce Orchestrator, an 8B model that achieves higher accuracy at lower cost than previous tool-use agents while aligning with user preferences on which tools are to be used for a given query. On HLE, Orchestrator achieves a score of 37.1%, outperforming GPT-5 (35.1%) while being 2.5x more efficient. On tau2-Bench and FRAMES, Orchestrator surpasses GPT-5 by a wide margin while using only about 30% of the cost. Extensive analysis shows that Orchestrator achieves the best trade-off between performance and cost under multiple metrics, and generalizes robustly to unseen tools. These results demonstrate that composing diverse tools with a lightweight orchestration model is both more efficient and more effective than existing methods, paving the way for practical and scalable tool-augmented reasoning systems.</p> + https://arxiv.org/abs/2511.21689 + Wed, 26 Nov 2025 18:59:46 +0000 - Controllable Layer Decomposition for Reversible Multi-Layer Image Generation - https://arxiv.org/abs/2511.16249 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.16249.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Zihao Liu, Zunnan Xu, Shi Shu, Jun Zhou, Ruicheng Zhang, Zhenchao Tang, Xiu Li</p><p><b>Upvotes:</b> 8</p><p><b>Summary:</b> This work presents Controllable Layer Decomposition (CLD), a method for achieving fine-grained and controllable multi-layer separation of raster images. In practical workflows, designers typically generate and edit each RGBA layer independently before compositing them into a final raster image. However, this process is irreversible: once composited, layer-level editing is no longer possible. Existing methods commonly rely on image matting and inpainting, but remain limited in controllability and segmentation precision. To address these challenges, we propose two key modules: LayerDecompose-DiT (LD-DiT), which decouples image elements into distinct layers and enables fine-grained control; and Multi-Layer Conditional Adapter (MLCA), which injects target image information into multi-layer tokens to achieve precise conditional generation. To enable a comprehensive evaluation, we build a new benchmark and introduce tailored evaluation metrics. Experimental results show that CLD consistently outperforms existing methods in both decomposition quality and controllability. Furthermore, the separated layers produced by CLD can be directly manipulated in commonly used design tools such as PowerPoint, highlighting its practical value and applicability in real-world creative workflows.</p> - https://arxiv.org/abs/2511.16249 - Thu, 20 Nov 2025 11:27:21 +0000 + DualVLA: Building a Generalizable Embodied Agent via Partial Decoupling of Reasoning and Action + https://arxiv.org/abs/2511.22134 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.22134.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Zhen Fang, Zhuoyang Liu, Jiaming Liu, Hao Chen, Yu Zeng, Shiting Huang, Zehui Chen, Lin Chen, Shanghang Zhang, Feng Zhao</p><p><b>Upvotes:</b> 21</p><p><b>Summary:</b> To build a generalizable Vision-Language-Action (VLA) model with strong reasoning ability, a common strategy is to first train a specialist VLA on robot demonstrations to acquire reliable manipulation skills, and then incorporate mixed annotated robot data together with multimodal data to restore broader reasoning capabilities. However, we observe that the resulting reasoning VLA often suffers from degraded action performance compared to the specialist model before fine-tuning, a phenomenon we refer to as action degeneration. To address this issue, we propose DualVLA, which enhances action performance through carefully designed post-training while still preserving reasoning capability. We first introduce a dual-layer data pruning method that removes redundant embodied reasoning, preventing it from adversely influencing action learning. To further strengthen action generation, we design a dual-teacher adaptive distillation strategy that assigns different supervision signals to different data domains while maintaining reasoning ability. To fill the evaluation gap for generalist VLAs, we also propose VLA Score, which decouples VLA capability into reasoning, intention, action, and alignment dimensions for a more fine-grained assessment. Experiments show that DualVLA achieves an average success rate of 61.0 in SimplerEnv and an average score of 65.4 across eight competitive multimodal benchmarks, demonstrating a stronger balance between precise action execution and multimodal understanding. Project Website: https://costaliya.github.io/DualVLA/.</p> + https://arxiv.org/abs/2511.22134 + Thu, 27 Nov 2025 06:03:53 +0000 - Upsample Anything: A Simple and Hard to Beat Baseline for Feature Upsampling - https://arxiv.org/abs/2511.16301 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.16301.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Minseok Seo, Mark Hamilton, Changick Kim</p><p><b>Upvotes:</b> 6</p><p><b>Summary:</b> We present Upsample Anything, a lightweight test-time optimization (TTO) framework that restores low-resolution features to high-resolution, pixel-wise outputs without any training. Although Vision Foundation Models demonstrate strong generalization across diverse downstream tasks, their representations are typically downsampled by 14x/16x (e.g., ViT), which limits their direct use in pixel-level applications. Existing feature upsampling approaches depend on dataset-specific retraining or heavy implicit optimization, restricting scalability and generalization. Upsample Anything addresses these issues through a simple per-image optimization that learns an anisotropic Gaussian kernel combining spatial and range cues, effectively bridging Gaussian Splatting and Joint Bilateral Upsampling. The learned kernel acts as a universal, edge-aware operator that transfers seamlessly across architectures and modalities, enabling precise high-resolution reconstruction of features, depth, or probability maps. It runs in only approx0.419 s per 224x224 image and achieves state-of-the-art performance on semantic segmentation, depth estimation, and both depth and probability map upsampling. Project page: https://seominseok0429.github.io/Upsample-Anything/{https://seominseok0429.github.io/Upsample-Anything/}</p> - https://arxiv.org/abs/2511.16301 - Thu, 20 Nov 2025 12:27:53 +0000 + RefineBench: Evaluating Refinement Capability of Language Models via Checklists + https://arxiv.org/abs/2511.22173 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.22173.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Young-Jun Lee, Seungone Kim, Byung-Kwan Lee, Minkyeong Moon, Yechan Hwang, Jong Myoung Kim, Graham Neubig, Sean Welleck, Ho-Jin Choi</p><p><b>Upvotes:</b> 12</p><p><b>Summary:</b> Can language models (LMs) self-refine their own responses? This question is increasingly relevant as a wide range of real-world user interactions involve refinement requests. However, prior studies have largely tested LMs' refinement abilities on verifiable tasks such as competition math or symbolic reasoning with simplified scaffolds, whereas users often pose open-ended queries and provide varying degrees of feedback on what they desire. The recent advent of reasoning models that exhibit self-reflection patterns in their chains-of-thought further motivates this question. To analyze this, we introduce RefineBench, a benchmark of 1,000 challenging problems across 11 domains paired with a checklist-based evaluation framework. We evaluate two refinement modes: (1) guided refinement, where an LM is provided natural language feedback, and (2) self-refinement, where LMs attempt to improve without guidance. In the self-refinement setting, even frontier LMs such as Gemini 2.5 Pro and GPT-5 achieve modest baseline scores of 31.3% and 29.1%, respectively, and most models fail to consistently improve across iterations (e.g., Gemini-2.5-Pro gains only +1.8%, while DeepSeek-R1 declines by -0.1%). By contrast, in guided refinement, both proprietary LMs and large open-weight LMs (>70B) can leverage targeted feedback to refine responses to near-perfect levels within five turns. These findings suggest that frontier LMs require breakthroughs to self-refine their incorrect responses, and that RefineBench provides a valuable testbed for tracking progress.</p> + https://arxiv.org/abs/2511.22173 + Thu, 27 Nov 2025 07:20:52 +0000 - OpenMMReasoner: Pushing the Frontiers for Multimodal Reasoning with an Open and General Recipe - https://arxiv.org/abs/2511.16334 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.16334.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Kaichen Zhang, Keming Wu, Zuhao Yang, Kairui Hu, Bin Wang, Ziwei Liu, Xingxuan Li, Lidong Bing</p><p><b>Upvotes:</b> 86</p><p><b>Summary:</b> Recent advancements in large reasoning models have fueled growing interest in extending such capabilities to multimodal domains. However, despite notable progress in visual reasoning, the lack of transparent and reproducible data curation and training strategies remains a major barrier to scalable research. In this work, we introduce OpenMMReasoner, a fully transparent two-stage recipe for multimodal reasoning spanning supervised fine-tuning (SFT) and reinforcement learning (RL). In the SFT stage, we construct an 874K-sample cold-start dataset with rigorous step-by-step validation, providing a strong foundation for reasoning capabilities. The subsequent RL stage leverages a 74K-sample dataset across diverse domains to further sharpen and stabilize these abilities, resulting in a more robust and efficient learning process. Extensive evaluations demonstrate that our training recipe not only surpasses strong baselines but also highlights the critical role of data quality and training design in shaping multimodal reasoning performance. Notably, our method achieves a 11.6% improvement over the Qwen2.5-VL-7B-Instruct baseline across nine multimodal reasoning benchmarks, establishing a solid empirical foundation for future large-scale multimodal reasoning research. We open-sourced all our codes, pipeline, and data at https://github.com/EvolvingLMMs-Lab/OpenMMReasoner.</p> - https://arxiv.org/abs/2511.16334 - Thu, 20 Nov 2025 13:11:45 +0000 + Adversarial Flow Models + https://arxiv.org/abs/2511.22475 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.22475.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Shanchuan Lin, Ceyuan Yang, Zhijie Lin, Hao Chen, Haoqi Fan</p><p><b>Upvotes:</b> 20</p><p><b>Summary:</b> We present adversarial flow models, a class of generative models that unifies adversarial models and flow models. Our method supports native one-step or multi-step generation and is trained using the adversarial objective. Unlike traditional GANs, where the generator learns an arbitrary transport plan between the noise and the data distributions, our generator learns a deterministic noise-to-data mapping, which is the same optimal transport as in flow-matching models. This significantly stabilizes adversarial training. Also, unlike consistency-based methods, our model directly learns one-step or few-step generation without needing to learn the intermediate timesteps of the probability flow for propagation. This saves model capacity, reduces training iterations, and avoids error accumulation. Under the same 1NFE setting on ImageNet-256px, our B/2 model approaches the performance of consistency-based XL/2 models, while our XL/2 model creates a new best FID of 2.38. We additionally show the possibility of end-to-end training of 56-layer and 112-layer models through depth repetition without any intermediate supervision, and achieve FIDs of 2.08 and 1.94 using a single forward pass, surpassing their 2NFE and 4NFE counterparts.</p> + https://arxiv.org/abs/2511.22475 + Thu, 27 Nov 2025 14:04:08 +0000 - AICC: Parse HTML Finer, Make Models Better -- A 7.3T AI-Ready Corpus Built by a Model-Based HTML Parser - https://arxiv.org/abs/2511.16397 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.16397.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Ren Ma, Jiantao Qiu, Chao Xu, Pei Chu, Kaiwen Liu, Pengli Ren, Yuan Qu, Jiahui Peng, Linfeng Hou, Mengjie Liu, Lindong Lu, Wenchang Ning, Jia Yu, Rui Min, Jin Shi, Haojiong Chen, Peng Zhang, Wenjian Zhang, Qian Jiang, Zengjie Hu, Guoqiang Yang, Zhenxiang Li, Fukai Shang, Zhongying Tu, Wentao Zhang, Dahua Lin, Conghui He</p><p><b>Upvotes:</b> 6</p><p><b>Summary:</b> While web data quality is crucial for large language models, most curation efforts focus on filtering and deduplication,treating HTML-to-text extraction as a fixed pre-processing step. Existing web corpora rely on heuristic-based extractors like Trafilatura, which struggle to preserve document structure and frequently corrupt structured elements such as formulas, codes, and tables. We hypothesize that improving extraction quality can be as impactful as aggressive filtering strategies for downstream performance. We introduce MinerU-HTML, a novel extraction pipeline that reformulates content extraction as a sequence labeling problem solved by a 0.6B-parameter language model. Unlike text-density heuristics, MinerU-HTML leverages semantic understanding and employs a two-stage formatting pipeline that explicitly categorizes semantic elements before converting to Markdown. Crucially, its model-based approach is inherently scalable, whereas heuristic methods offer limited improvement pathways. On MainWebBench, our benchmark of 7,887 annotated web pages, MinerU-HTML achieves 81.8\% ROUGE-N F1 compared to Trafilatura's 63.6\%, with exceptional structured element preservation (90.9\% for code blocks, 94.0\% for formulas). Using MinerU-HTML, we construct AICC (AI-ready Common Crawl), a 7.3-trillion token multilingual corpus from two Common Crawl snapshots. In controlled pretraining experiments where AICC and Trafilatura-extracted TfCC undergo identical filtering, models trained on AICC (62B tokens) achieve 50.8\% average accuracy across 13 benchmarks, outperforming TfCC by 1.08pp-providing direct evidence that extraction quality significantly impacts model capabilities. AICC also surpasses RefinedWeb and FineWeb on key benchmarks. We publicly release MainWebBench, MinerU-HTML, and AICC, demonstrating that HTML extraction is a critical, often underestimated component of web corpus construction.</p> - https://arxiv.org/abs/2511.16397 - Thu, 20 Nov 2025 14:15:23 +0000 + DeepSeekMath-V2: Towards Self-Verifiable Mathematical Reasoning + https://arxiv.org/abs/2511.22570 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.22570.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Zhihong Shao, Yuxiang Luo, Chengda Lu, Z. Z. Ren, Jiewen Hu, Tian Ye, Zhibin Gou, Shirong Ma, Xiaokang Zhang</p><p><b>Upvotes:</b> 65</p><p><b>Summary:</b> Large language models have made significant progress in mathematical reasoning, which serves as an important testbed for AI and could impact scientific research if further advanced. By scaling reasoning with reinforcement learning that rewards correct final answers, LLMs have improved from poor performance to saturating quantitative reasoning competitions like AIME and HMMT in one year. However, this approach faces fundamental limitations. Pursuing higher final answer accuracy doesn't address a key issue: correct answers don't guarantee correct reasoning. Moreover, many mathematical tasks like theorem proving require rigorous step-by-step derivation rather than numerical answers, making final answer rewards inapplicable. To push the limits of deep reasoning, we believe it is necessary to verify the comprehensiveness and rigor of mathematical reasoning. Self-verification is particularly important for scaling test-time compute, especially for open problems without known solutions. Towards self-verifiable mathematical reasoning, we investigate how to train an accurate and faithful LLM-based verifier for theorem proving. We then train a proof generator using the verifier as the reward model, and incentivize the generator to identify and resolve as many issues as possible in their own proofs before finalizing them. To maintain the generation-verification gap as the generator becomes stronger, we propose to scale verification compute to automatically label new hard-to-verify proofs, creating training data to further improve the verifier. Our resulting model, DeepSeekMath-V2, demonstrates strong theorem-proving capabilities, achieving gold-level scores on IMO 2025 and CMO 2024 and a near-perfect 118/120 on Putnam 2024 with scaled test-time compute.</p> + https://arxiv.org/abs/2511.22570 + Thu, 27 Nov 2025 16:01:22 +0000 - Cognitive Foundations for Reasoning and Their Manifestation in LLMs - https://arxiv.org/abs/2511.16660 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.16660.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Priyanka Kargupta, Shuyue Stella Li, Haocheng Wang, Jinu Lee, Shan Chen, Orevaoghene Ahia, Dean Light, Thomas L. Griffiths, Max Kleiman-Weiner, Jiawei Han, Asli Celikyilmaz, Yulia Tsvetkov</p><p><b>Upvotes:</b> 8</p><p><b>Summary:</b> Large language models (LLMs) solve complex problems yet fail on simpler variants, suggesting they achieve correct outputs through mechanisms fundamentally different from human reasoning. To understand this gap, we synthesize cognitive science research into a taxonomy of 28 cognitive elements spanning reasoning invariants, meta-cognitive controls, representations for organizing reasoning & knowledge, and transformation operations. We introduce a fine-grained evaluation framework and conduct the first large-scale empirical analysis of 192K traces from 18 models across text, vision, and audio, complemented by 54 human think-aloud traces, which we make publicly available. We find that models under-utilize cognitive elements correlated with success, narrowing to rigid sequential processing on ill-structured problems where diverse representations and meta-cognitive monitoring are critical. Human traces show more abstraction and conceptual processing, while models default to surface-level enumeration. Meta-analysis of 1.6K LLM reasoning papers reveals the research community concentrates on easily quantifiable elements (sequential organization: 55%, decomposition: 60%) but neglecting meta-cognitive controls (self-awareness: 16%) that correlate with success. Models possess behavioral repertoires associated with success but fail to deploy them spontaneously. Leveraging these patterns, we develop test-time reasoning guidance that automatically scaffold successful structures, improving performance by up to 66.7% on complex problems. By establishing a shared vocabulary between cognitive science and LLM research, our framework enables systematic diagnosis of reasoning failures and principled development of models that reason through robust cognitive mechanisms rather than spurious shortcuts, while providing tools to test theories of human cognition at scale.</p> - https://arxiv.org/abs/2511.16660 - Thu, 20 Nov 2025 18:59:00 +0000 + Revisiting the Necessity of Lengthy Chain-of-Thought in Vision-centric Reasoning Generalization + https://arxiv.org/abs/2511.22586 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.22586.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Yifan Du, Kun Zhou, Yingqian Min, Yue Ling, Wayne Xin Zhao, Youbin Wu</p><p><b>Upvotes:</b> 6</p><p><b>Summary:</b> We study how different Chain-of-Thought (CoT) designs affect the acquisition of the generalizable visual reasoning ability in vision-language models (VLMs). While CoT data, especially long or visual CoT such as "think with image", has been widely used to supervise intermediate reasoning, it remains unclear why specific CoT designs help and which ones truly support generalizable reasoning. To systematically evaluate this, we focus on a controlled maze-solving benchmark where reasoning rules are fully visual, difficulty can be tuned by grid size, and all the intermediate steps can be automatically generated. Using Qwen2.5-VL-7B under a standard SFT-then-RL pipeline, we compare three representative CoT formats: Language CoT, Grounding CoT (with spatial coordinate trajectories), and Visual CoT (with image manipulations). Our experiments reveal that visual and longer CoT mainly accelerate convergence but do not lift the final performance ceiling; concise CoT containing only essential grounding steps outperforms longer traces; and, strikingly, CoT retaining only the minimal grounding results generalizes best across different maze sizes. We further validate these insights on other vision-centric tasks. These findings highlight a "short is long" effect and provide practical guidance for constructing more generalizable SFT datasets for visual reasoning.</p> + https://arxiv.org/abs/2511.22586 + Thu, 27 Nov 2025 16:19:34 +0000 - SAM 3: Segment Anything with Concepts - https://arxiv.org/abs/2511.16719 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.16719.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Nicolas Carion, Laura Gustafson, Yuan-Ting Hu, Shoubhik Debnath, Ronghang Hu, Didac Suris, Chaitanya Ryali, Kalyan Vasudev Alwala, Haitham Khedr, Andrew Huang, Jie Lei, Tengyu Ma, Baishan Guo, Arpit Kalla, Markus Marks, Joseph Greer, Meng Wang, Peize Sun, Roman Rädle, Triantafyllos Afouras, Effrosyni Mavroudi, Katherine Xu, Tsung-Han Wu, Yu Zhou, Liliane Momeni, Rishi Hazra, Shuangrui Ding, Sagar Vaze, Francois Porcher, Feng Li, Siyuan Li, Aishwarya Kamath, Ho Kei Cheng, Piotr Dollár, Nikhila Ravi, Kate Saenko, Pengchuan Zhang, Christoph Feichtenhofer</p><p><b>Upvotes:</b> 94</p><p><b>Summary:</b> We present Segment Anything Model (SAM) 3, a unified model that detects, segments, and tracks objects in images and videos based on concept prompts, which we define as either short noun phrases (e.g., "yellow school bus"), image exemplars, or a combination of both. Promptable Concept Segmentation (PCS) takes such prompts and returns segmentation masks and unique identities for all matching object instances. To advance PCS, we build a scalable data engine that produces a high-quality dataset with 4M unique concept labels, including hard negatives, across images and videos. Our model consists of an image-level detector and a memory-based video tracker that share a single backbone. Recognition and localization are decoupled with a presence head, which boosts detection accuracy. SAM 3 doubles the accuracy of existing systems in both image and video PCS, and improves previous SAM capabilities on visual segmentation tasks. We open source SAM 3 along with our new Segment Anything with Concepts (SA-Co) benchmark for promptable concept segmentation.</p> - https://arxiv.org/abs/2511.16719 - Thu, 20 Nov 2025 18:59:56 +0000 + MG-Nav: Dual-Scale Visual Navigation via Sparse Spatial Memory + https://arxiv.org/abs/2511.22609 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.22609.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Bo Wang, Jiehong Lin, Chenzhi Liu, Xinting Hu, Yifei Yu, Tianjia Liu, Zhongrui Wang, Xiaojuan Qi</p><p><b>Upvotes:</b> 47</p><p><b>Summary:</b> We present MG-Nav (Memory-Guided Navigation), a dual-scale framework for zero-shot visual navigation that unifies global memory-guided planning with local geometry-enhanced control. At its core is the Sparse Spatial Memory Graph (SMG), a compact, region-centric memory where each node aggregates multi-view keyframe and object semantics, capturing both appearance and spatial structure while preserving viewpoint diversity. At the global level, the agent is localized on SMG and a goal-conditioned node path is planned via an image-to-instance hybrid retrieval, producing a sequence of reachable waypoints for long-horizon guidance. At the local level, a navigation foundation policy executes these waypoints in point-goal mode with obstacle-aware control, and switches to image-goal mode when navigating from the final node towards the visual target. To further enhance viewpoint alignment and goal recognition, we introduce VGGT-adapter, a lightweight geometric module built on the pre-trained VGGT model, which aligns observation and goal features in a shared 3D-aware space. MG-Nav operates global planning and local control at different frequencies, using periodic re-localization to correct errors. Experiments on HM3D Instance-Image-Goal and MP3D Image-Goal benchmarks demonstrate that MG-Nav achieves state-of-the-art zero-shot performance and remains robust under dynamic rearrangements and unseen scene conditions.</p> + https://arxiv.org/abs/2511.22609 + Thu, 27 Nov 2025 16:43:21 +0000 - WorldGen: From Text to Traversable and Interactive 3D Worlds - https://arxiv.org/abs/2511.16825 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.16825.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Dilin Wang, Hyunyoung Jung, Tom Monnier, Kihyuk Sohn, Chuhang Zou, Xiaoyu Xiang, Yu-Ying Yeh, Di Liu, Zixuan Huang, Thu Nguyen-Phuoc, Yuchen Fan, Sergiu Oprea, Ziyan Wang, Roman Shapovalov, Nikolaos Sarafianos, Thibault Groueix, Antoine Toisoul, Prithviraj Dhar, Xiao Chu, Minghao Chen, Geon Yeong Park, Mahima Gupta, Yassir Azziz, Rakesh Ranjan, Andrea Vedaldi</p><p><b>Upvotes:</b> 18</p><p><b>Summary:</b> We introduce WorldGen, a system that enables the automatic creation of large-scale, interactive 3D worlds directly from text prompts. Our approach transforms natural language descriptions into traversable, fully textured environments that can be immediately explored or edited within standard game engines. By combining LLM-driven scene layout reasoning, procedural generation, diffusion-based 3D generation, and object-aware scene decomposition, WorldGen bridges the gap between creative intent and functional virtual spaces, allowing creators to design coherent, navigable worlds without manual modeling or specialized 3D expertise. The system is fully modular and supports fine-grained control over layout, scale, and style, producing worlds that are geometrically consistent, visually rich, and efficient to render in real time. This work represents a step towards accessible, generative world-building at scale, advancing the frontier of 3D generative AI for applications in gaming, simulation, and immersive social environments.</p> - https://arxiv.org/abs/2511.16825 - Thu, 20 Nov 2025 22:13:18 +0000 + REASONEDIT: Towards Reasoning-Enhanced Image Editing Models + https://arxiv.org/abs/2511.22625 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.22625.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Fukun Yin, Shiyu Liu, Yucheng Han, Zhibo Wang, Peng Xing, Rui Wang, Wei Cheng, Yingming Wang, Aojie Li, Zixin Yin, Pengtao Chen, Xiangyu Zhang, Daxin Jiang, Xianfang Zeng, Gang Yu</p><p><b>Upvotes:</b> 45</p><p><b>Summary:</b> Recent advances in image editing models have shown remarkable progress. A common architectural design couples a multimodal large language model (MLLM) encoder with a diffusion decoder, as seen in systems such as Step1X-Edit and Qwen-Image-Edit, where the MLLM encodes both the reference image and the instruction but remains frozen during training. In this work, we demonstrate that unlocking the reasoning capabilities of MLLM can further push the boundaries of editing models. Specifically, we explore two reasoning mechanisms, thinking and reflection, which enhance instruction understanding and editing accuracy. Based on that, our proposed framework enables image editing in a thinking-editing-reflection loop: the thinking mechanism leverages the world knowledge of MLLM to interpret abstract instructions, while the reflection reviews editing results, automatically corrects unintended manipulations, and identifies the stopping round. Extensive experiments demonstrate that our reasoning approach achieves significant performance gains, with improvements of ImgEdit (+4.3%), GEdit (+4.7%), and Kris (+8.2%) when initializing our DiT from the Step1X-Edit (ReasonEdit-S), and also outperforms previous open-source methods on both GEdit and Kris when integrated with Qwen-Image-Edit (ReasonEdit-Q).</p> + https://arxiv.org/abs/2511.22625 + Thu, 27 Nov 2025 17:02:48 +0000 - OmniScientist: Toward a Co-evolving Ecosystem of Human and AI Scientists - https://arxiv.org/abs/2511.16931 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.16931.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Chenyang Shao, Dehao Huang, Yu Li, Keyu Zhao, Weiquan Lin, Yining Zhang, Qingbin Zeng, Zhiyu Chen, Tianxing Li, Yifei Huang, Taozhong Wu, Xinyang Liu, Ruotong Zhao, Mengsheng Zhao, Xuhua Zhang, Yue Wang, Yuanyi Zhen, Fengli Xu, Yong Li, Tie-Yan Liu</p><p><b>Upvotes:</b> 6</p><p><b>Summary:</b> With the rapid development of Large Language Models (LLMs), AI agents have demonstrated increasing proficiency in scientific tasks, ranging from hypothesis generation and experimental design to manuscript writing. Such agent systems are commonly referred to as "AI Scientists." However, existing AI Scientists predominantly formulate scientific discovery as a standalone search or optimization problem, overlooking the fact that scientific research is inherently a social and collaborative endeavor. Real-world science relies on a complex scientific infrastructure composed of collaborative mechanisms, contribution attribution, peer review, and structured scientific knowledge networks. Due to the lack of modeling for these critical dimensions, current systems struggle to establish a genuine research ecosystem or interact deeply with the human scientific community. To bridge this gap, we introduce OmniScientist, a framework that explicitly encodes the underlying mechanisms of human research into the AI scientific workflow. OmniScientist not only achieves end-to-end automation across data foundation, literature review, research ideation, experiment automation, scientific writing, and peer review, but also provides comprehensive infrastructural support by simulating the human scientific system, comprising: (1) a structured knowledge system built upon citation networks and conceptual correlations; (2) a collaborative research protocol (OSP), which enables seamless multi-agent collaboration and human researcher participation; and (3) an open evaluation platform (ScienceArena) based on blind pairwise user voting and Elo rankings. This infrastructure empowers agents to not only comprehend and leverage human knowledge systems but also to collaborate and co-evolve, fostering a sustainable and scalable innovation ecosystem.</p> - https://arxiv.org/abs/2511.16931 - Fri, 21 Nov 2025 03:55:19 +0000 + Geometrically-Constrained Agent for Spatial Reasoning + https://arxiv.org/abs/2511.22659 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.22659.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Zeren Chen, Xiaoya Lu, Zhijie Zheng, Pengrui Li, Lehan He, Yijin Zhou, Jing Shao, Bohan Zhuang, Lu Sheng</p><p><b>Upvotes:</b> 38</p><p><b>Summary:</b> Vision Language Models (VLMs) exhibit a fundamental semantic-to-geometric gap in spatial reasoning: they excel at qualitative semantic inference but their reasoning operates within a lossy semantic space, misaligned with high-fidelity geometry. Current paradigms fail to bridge this gap. Training-based methods suffer from an ``oracle paradox,'' learning flawed spatial logic from imperfect oracles. Tool-integrated methods constrain the final computation but critically leave the VLM's planning process unconstrained, resulting in geometrically flawed plans. In this work, we propose Geometrically-Constrained Agent (GCA), a training-free agentic paradigm that resolves this gap by introducing a formal task constraint. Specifically, we strategically decouples the VLM's role into two stages. First, acting as a semantic analyst, the VLM translates the user's ambiguous query into the formal, verifiable task constraint, which defines the reference frame and objective. Second, acting as a task solver, the VLM generates and executes tool calls strictly within the deterministic bounds defined by the constraint. This geometrically-constrained reasoning strategy successfully resolve the semantic-to-geometric gap, yielding a robust and verifiable reasoning pathway for spatial reasoning. Comprehensive experiments demonstrate that GCA achieves SOTA performance on multiple spatial reasoning benchmarks, surpassing existing training-based and tool-integrated methods by ~27%. Please see our homepage at https://gca-spatial-reasoning.github.io.</p> + https://arxiv.org/abs/2511.22659 + Thu, 27 Nov 2025 17:50:37 +0000 - Budget-Aware Tool-Use Enables Effective Agent Scaling - https://arxiv.org/abs/2511.17006 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.17006.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Tengxiao Liu, Zifeng Wang, Jin Miao, I-Hung Hsu, Jun Yan, Jiefeng Chen, Rujun Han, Fangyuan Xu, Yanfei Chen, Ke Jiang, Samira Daruki, Yi Liang, William Yang Wang, Tomas Pfister, Chen-Yu Lee</p><p><b>Upvotes:</b> 24</p><p><b>Summary:</b> Scaling test-time computation improves performance across different tasks on large language models (LLMs), which has also been extended to tool-augmented agents. For these agents, scaling involves not only "thinking" in tokens but also "acting" via tool calls. The number of tool calls directly bounds the agent's interaction with the external environment. However, we find that simply granting agents a larger tool-call budget fails to improve performance, as they lack "budget awareness" and quickly hit a performance ceiling. To address this, we study how to scale such agents effectively under explicit tool-call budgets, focusing on web search agents. We first introduce the Budget Tracker, a lightweight plug-in that provides the agent with continuous budget awareness, enabling simple yet effective scaling. We further develop BATS (Budget Aware Test-time Scaling), an advanced framework that leverages this awareness to dynamically adapt its planning and verification strategy, deciding whether to "dig deeper" on a promising lead or "pivot" to new paths based on remaining resources. To analyze cost-performance scaling in a controlled manner, we formalize a unified cost metric that jointly accounts for token and tool consumption. We provide the first systematic study on budget-constrained agents, showing that budget-aware methods produce more favorable scaling curves and push the cost-performance Pareto frontier. Our work offers empirical insights toward a more transparent and principled understanding of scaling in tool-augmented agents.</p> - https://arxiv.org/abs/2511.17006 - Fri, 21 Nov 2025 07:18:55 +0000 + Architecture Decoupling Is Not All You Need For Unified Multimodal Model + https://arxiv.org/abs/2511.22663 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.22663.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Dian Zheng, Manyuan Zhang, Hongyu Li, Kai Zou, Hongbo Liu, Ziyu Guo, Kaituo Feng, Yexin Liu, Ying Luo, Yan Feng, Peng Pei, Xunliang Cai, Hongsheng Li</p><p><b>Upvotes:</b> 28</p><p><b>Summary:</b> Unified multimodal models for image generation and understanding represent a significant step toward AGI and have attracted widespread attention from researchers. The main challenge of this task lies in the difficulty in establishing an optimal training paradigm due to inherent conflicting targets in understanding and generation tasks. To alleviate these conflicts and pursue higher performance, many researchers adopt varying degrees of model decoupling (e.g., Double image encoders, MOE/MOT architecture, or frozen MLLM). However, excessive model decoupling can lead to the loss of interleave generation ability, undermining the original intent of unified models. In this work, we aim to explore how to mitigate task conflicts without resorting to model decoupling. Firstly, we analyze why decoupling alleviates conflicts by studying the cross-modal attention behavior of models. We observe that model decoupling essentially drives models toward task-specific multimodal interaction patterns, as seen in Qwen-VL and HunyuanImage, and that the more thorough the decoupling, the more consistent the behavior becomes. Motivated by this observation, we propose Attention Interaction Alignment (AIA) loss, which explicitly learns Task-Specific multimodal interaction patterns during training. To demonstrate the generalizability of our AIA loss, we apply it to Emu3 and Janus-Pro during SFT and post-training stage respectively. Without bells and whistles, AIA not only refines cross-modal attention patterns, but also boosts both generation and understanding performance.</p> + https://arxiv.org/abs/2511.22663 + Thu, 27 Nov 2025 17:55:25 +0000 - Diversity Has Always Been There in Your Visual Autoregressive Models - https://arxiv.org/abs/2511.17074 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.17074.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Tong Wang, Guanyu Yang, Nian Liu, Kai Wang, Yaxing Wang, Abdelrahman M Shaker, Salman Khan, Fahad Shahbaz Khan, Senmao Li</p><p><b>Upvotes:</b> 6</p><p><b>Summary:</b> Visual Autoregressive (VAR) models have recently garnered significant attention for their innovative next-scale prediction paradigm, offering notable advantages in both inference efficiency and image quality compared to traditional multi-step autoregressive (AR) and diffusion models. However, despite their efficiency, VAR models often suffer from the diversity collapse i.e., a reduction in output variability, analogous to that observed in few-step distilled diffusion models. In this paper, we introduce DiverseVAR, a simple yet effective approach that restores the generative diversity of VAR models without requiring any additional training. Our analysis reveals the pivotal component of the feature map as a key factor governing diversity formation at early scales. By suppressing the pivotal component in the model input and amplifying it in the model output, DiverseVAR effectively unlocks the inherent generative potential of VAR models while preserving high-fidelity synthesis. Empirical results demonstrate that our approach substantially enhances generative diversity with only neglectable performance influences. Our code will be publicly released at https://github.com/wangtong627/DiverseVAR.</p> - https://arxiv.org/abs/2511.17074 - Fri, 21 Nov 2025 09:24:09 +0000 + Decoupled DMD: CFG Augmentation as the Spear, Distribution Matching as the Shield + https://arxiv.org/abs/2511.22677 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.22677.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Dongyang Liu, Peng Gao, David Liu, Ruoyi Du, Zhen Li, Qilong Wu, Xin Jin, Sihan Cao, Shifeng Zhang, Hongsheng Li, Steven Hoi</p><p><b>Upvotes:</b> 20</p><p><b>Summary:</b> Diffusion model distillation has emerged as a powerful technique for creating efficient few-step and single-step generators. Among these, Distribution Matching Distillation (DMD) and its variants stand out for their impressive performance, which is widely attributed to their core mechanism of matching the student's output distribution to that of a pre-trained teacher model. In this work, we challenge this conventional understanding. Through a rigorous decomposition of the DMD training objective, we reveal that in complex tasks like text-to-image generation, where CFG is typically required for desirable few-step performance, the primary driver of few-step distillation is not distribution matching, but a previously overlooked component we identify as CFG Augmentation (CA). We demonstrate that this term acts as the core ``engine'' of distillation, while the Distribution Matching (DM) term functions as a ``regularizer'' that ensures training stability and mitigates artifacts. We further validate this decoupling by demonstrating that while the DM term is a highly effective regularizer, it is not unique; simpler non-parametric constraints or GAN-based objectives can serve the same stabilizing function, albeit with different trade-offs. This decoupling of labor motivates a more principled analysis of the properties of both terms, leading to a more systematic and in-depth understanding. This new understanding further enables us to propose principled modifications to the distillation process, such as decoupling the noise schedules for the engine and the regularizer, leading to further performance gains. Notably, our method has been adopted by the Z-Image ( https://github.com/Tongyi-MAI/Z-Image ) project to develop a top-tier 8-step image generation model, empirically validating the generalization and robustness of our findings.</p> + https://arxiv.org/abs/2511.22677 + Thu, 27 Nov 2025 18:24:28 +0000 - VLA-4D: Embedding 4D Awareness into Vision-Language-Action Models for SpatioTemporally Coherent Robotic Manipulation - https://arxiv.org/abs/2511.17199 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.17199.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Hanyu Zhou, Chuanhao Ma, Gim Hee Lee</p><p><b>Upvotes:</b> 6</p><p><b>Summary:</b> Vision-language-action (VLA) models show potential for general robotic tasks, but remain challenging in spatiotemporally coherent manipulation, which requires fine-grained representations. Typically, existing methods embed 3D positions into visual representations to enhance the spatial precision of actions. However, these methods struggle to achieve temporally coherent control over action execution. In this work, we propose VLA-4D, a general VLA model with 4D awareness for spatiotemporally coherent robotic manipulation. Our model is guided by two key designs: 1) 4D-aware visual representation. We extract visual features, embed 1D time into 3D positions for 4D embeddings, and fuse them into a unified visual representation via a cross-attention mechanism. 2) Spatiotemporal action representation. We extend conventional spatial action representations with temporal information to enable the spatiotemporal planning, and align the multimodal representations into the LLM for spatiotemporal action prediction. Within this unified framework, the designed visual and action representations jointly make robotic manipulation spatially-smooth and temporally-coherent. In addition, we extend the VLA dataset with temporal action annotations for fine-tuning our model. Extensive experiments have been conducted to verify the superiority of our method across different tasks of robotic manipulation.</p> - https://arxiv.org/abs/2511.17199 - Fri, 21 Nov 2025 12:26:30 +0000 + Z-Image: An Efficient Image Generation Foundation Model with Single-Stream Diffusion Transformer + https://arxiv.org/abs/2511.22699 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.22699.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Z-Image Team, Huanqia Cai, Sihan Cao, Ruoyi Du, Peng Gao, Steven Hoi, Shijie Huang, Zhaohui Hou, Dengyang Jiang, Xin Jin, Liangchen Li, Zhen Li, Zhong-Yu Li, David Liu, Dongyang Liu, Junhan Shi, Qilong Wu, Feng Yu, Chi Zhang, Shifeng Zhang, Shilin Zhou</p><p><b>Upvotes:</b> 152</p><p><b>Summary:</b> The landscape of high-performance image generation models is currently dominated by proprietary systems, such as Nano Banana Pro and Seedream 4.0. Leading open-source alternatives, including Qwen-Image, Hunyuan-Image-3.0 and FLUX.2, are characterized by massive parameter counts (20B to 80B), making them impractical for inference, and fine-tuning on consumer-grade hardware. To address this gap, we propose Z-Image, an efficient 6B-parameter foundation generative model built upon a Scalable Single-Stream Diffusion Transformer (S3-DiT) architecture that challenges the "scale-at-all-costs" paradigm. By systematically optimizing the entire model lifecycle -- from a curated data infrastructure to a streamlined training curriculum -- we complete the full training workflow in just 314K H800 GPU hours (approx. $630K). Our few-step distillation scheme with reward post-training further yields Z-Image-Turbo, offering both sub-second inference latency on an enterprise-grade H800 GPU and compatibility with consumer-grade hardware (<16GB VRAM). Additionally, our omni-pre-training paradigm also enables efficient training of Z-Image-Edit, an editing model with impressive instruction-following capabilities. Both qualitative and quantitative experiments demonstrate that our model achieves performance comparable to or surpassing that of leading competitors across various dimensions. Most notably, Z-Image exhibits exceptional capabilities in photorealistic image generation and bilingual text rendering, delivering results that rival top-tier commercial models, thereby demonstrating that state-of-the-art results are achievable with significantly reduced computational overhead. We publicly release our code, weights, and online demo to foster the development of accessible, budget-friendly, yet state-of-the-art generative models.</p> + https://arxiv.org/abs/2511.22699 + Thu, 27 Nov 2025 18:52:07 +0000 - Parrot: Persuasion and Agreement Robustness Rating of Output Truth -- A Sycophancy Robustness Benchmark for LLMs - https://arxiv.org/abs/2511.17220 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.17220.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Yusuf Çelebi, Mahmoud El Hussieni, Özay Ezerceli</p><p><b>Upvotes:</b> 15</p><p><b>Summary:</b> This study presents PARROT (Persuasion and Agreement Robustness Rating of Output Truth), a robustness focused framework designed to measure the degradation in accuracy that occurs under social pressure exerted on users through authority and persuasion in large language models (LLMs) the phenomenon of sycophancy (excessive conformity). PARROT (i) isolates causal effects by comparing the neutral version of the same question with an authoritatively false version using a double-blind evaluation, (ii) quantifies confidence shifts toward the correct and imposed false responses using log-likelihood-based calibration tracking, and (iii) systematically classifies failure modes (e.g., robust correct, sycophantic agreement, reinforced error, stubborn error, self-correction, etc.) using an eight-state behavioral taxonomy. We evaluated 22 models using 1,302 MMLU-style multiple-choice questions across 13 domains and domain-specific authority templates. Findings show marked heterogeneity: advanced models (e.g., GPT-5, GPT-4.1, Claude Sonnet 4.5) exhibit low "follow rates" (leq 11%, GPT-5: 4\%) and minimal accuracy loss, while older/smaller models show severe epistemic collapse (GPT-4: 80\%, Qwen 2.5-1.5B: 94\%). The danger is not limited to response changes; weak models reduce confidence in the correct response while increasing confidence in the imposed incorrect response. While international law and global knowledge at the domain level exhibit high fragility, elementary mathematics is relatively resilient. Consequently, we argue that the goal of "resistance to overfitting pressure" should be addressed as a primary objective alongside accuracy, harm avoidance, and privacy for safe deployment in the real world.</p> - https://arxiv.org/abs/2511.17220 - Fri, 21 Nov 2025 13:01:28 +0000 + World in a Frame: Understanding Culture Mixing as a New Challenge for Vision-Language Models + https://arxiv.org/abs/2511.22787 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.22787.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Eunsu Kim, Junyeong Park, Na Min An, Junseong Kim, Hitesh Laxmichand Patel, Jiho Jin, Julia Kruk, Amit Agarwal, Srikant Panda, Fenal Ashokbhai Ilasariya, Hyunjung Shim, Alice Oh</p><p><b>Upvotes:</b> 8</p><p><b>Summary:</b> In a globalized world, cultural elements from diverse origins frequently appear together within a single visual scene. We refer to these as culture mixing scenarios, yet how Large Vision-Language Models (LVLMs) perceive them remains underexplored. We investigate culture mixing as a critical challenge for LVLMs and examine how current models behave when cultural items from multiple regions appear together. To systematically analyze these behaviors, we construct CultureMix, a food Visual Question Answering (VQA) benchmark with 23k diffusion-generated, human-verified culture mixing images across four subtasks: (1) food-only, (2) food+food, (3) food+background, and (4) food+food+background. Evaluating 10 LVLMs, we find consistent failures to preserve individual cultural identities in mixed settings. Models show strong background reliance, with accuracy dropping 14% when cultural backgrounds are added to food-only baselines, and they produce inconsistent predictions for identical foods across different contexts. To address these limitations, we explore three robustness strategies. We find supervised fine-tuning using a diverse culture mixing dataset substantially improve model consistency and reduce background sensitivity. We call for increased attention to culture mixing scenarios as a critical step toward developing LVLMs capable of operating reliably in culturally diverse real-world environments.</p> + https://arxiv.org/abs/2511.22787 + Thu, 27 Nov 2025 22:23:08 +0000 - Loomis Painter: Reconstructing the Painting Process - https://arxiv.org/abs/2511.17344 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.17344.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Markus Pobitzer, Chang Liu, Chenyi Zhuang, Teng Long, Bin Ren, Nicu Sebe</p><p><b>Upvotes:</b> 15</p><p><b>Summary:</b> Step-by-step painting tutorials are vital for learning artistic techniques, but existing video resources (e.g., YouTube) lack interactivity and personalization. While recent generative models have advanced artistic image synthesis, they struggle to generalize across media and often show temporal or structural inconsistencies, hindering faithful reproduction of human creative workflows. To address this, we propose a unified framework for multi-media painting process generation with a semantics-driven style control mechanism that embeds multiple media into a diffusion models conditional space and uses cross-medium style augmentation. This enables consistent texture evolution and process transfer across styles. A reverse-painting training strategy further ensures smooth, human-aligned generation. We also build a large-scale dataset of real painting processes and evaluate cross-media consistency, temporal coherence, and final-image fidelity, achieving strong results on LPIPS, DINO, and CLIP metrics. Finally, our Perceptual Distance Profile (PDP) curve quantitatively models the creative sequence, i.e., composition, color blocking, and detail refinement, mirroring human artistic progression.</p> - https://arxiv.org/abs/2511.17344 - Fri, 21 Nov 2025 16:06:32 +0000 + Captain Safari: A World Engine + https://arxiv.org/abs/2511.22815 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.22815.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Yu-Cheng Chou, Xingrui Wang, Yitong Li, Jiahao Wang, Hanting Liu, Cihang Xie, Alan Yuille, Junfei Xiao</p><p><b>Upvotes:</b> 9</p><p><b>Summary:</b> World engines aim to synthesize long, 3D-consistent videos that support interactive exploration of a scene under user-controlled camera motion. However, existing systems struggle under aggressive 6-DoF trajectories and complex outdoor layouts: they lose long-range geometric coherence, deviate from the target path, or collapse into overly conservative motion. To this end, we introduce Captain Safari, a pose-conditioned world engine that generates videos by retrieving from a persistent world memory. Given a camera path, our method maintains a dynamic local memory and uses a retriever to fetch pose-aligned world tokens, which then condition video generation along the trajectory. This design enables the model to maintain stable 3D structure while accurately executing challenging camera maneuvers. To evaluate this setting, we curate OpenSafari, a new in-the-wild FPV dataset containing high-dynamic drone videos with verified camera trajectories, constructed through a multi-stage geometric and kinematic validation pipeline. Across video quality, 3D consistency, and trajectory following, Captain Safari substantially outperforms state-of-the-art camera-controlled generators. It reduces MEt3R from 0.3703 to 0.3690, improves AUC@30 from 0.181 to 0.200, and yields substantially lower FVD than all camera-controlled baselines. More importantly, in a 50-participant, 5-way human study where annotators select the best result among five anonymized models, 67.6% of preferences favor our method across all axes. Our results demonstrate that pose-conditioned world memory is a powerful mechanism for long-horizon, controllable video generation and provide OpenSafari as a challenging new benchmark for future world-engine research.</p> + https://arxiv.org/abs/2511.22815 + Fri, 28 Nov 2025 00:27:46 +0000 - Planning with Sketch-Guided Verification for Physics-Aware Video Generation - https://arxiv.org/abs/2511.17450 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.17450.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Yidong Huang, Zun Wang, Han Lin, Dong-Ki Kim, Shayegan Omidshafiei, Jaehong Yoon, Yue Zhang, Mohit Bansal</p><p><b>Upvotes:</b> 2</p><p><b>Summary:</b> Recent video generation approaches increasingly rely on planning intermediate control signals such as object trajectories to improve temporal coherence and motion fidelity. However, these methods mostly employ single-shot plans that are typically limited to simple motions, or iterative refinement which requires multiple calls to the video generator, incuring high computational cost. To overcome these limitations, we propose SketchVerify, a training-free, sketch-verification-based planning framework that improves motion planning quality with more dynamically coherent trajectories (i.e., physically plausible and instruction-consistent motions) prior to full video generation by introducing a test-time sampling and verification loop. Given a prompt and a reference image, our method predicts multiple candidate motion plans and ranks them using a vision-language verifier that jointly evaluates semantic alignment with the instruction and physical plausibility. To efficiently score candidate motion plans, we render each trajectory as a lightweight video sketch by compositing objects over a static background, which bypasses the need for expensive, repeated diffusion-based synthesis while achieving comparable performance. We iteratively refine the motion plan until a satisfactory one is identified, which is then passed to the trajectory-conditioned generator for final synthesis. Experiments on WorldModelBench and PhyWorldBench demonstrate that our method significantly improves motion quality, physical realism, and long-term consistency compared to competitive baselines while being substantially more efficient. Our ablation study further shows that scaling up the number of trajectory candidates consistently enhances overall performance.</p> - https://arxiv.org/abs/2511.17450 - Fri, 21 Nov 2025 17:48:02 +0000 + ORION: Teaching Language Models to Reason Efficiently in the Language of Thought + https://arxiv.org/abs/2511.22891 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.22891.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Kumar Tanmay, Kriti Aggarwal, Paul Pu Liang, Subhabrata Mukherjee</p><p><b>Upvotes:</b> 7</p><p><b>Summary:</b> Large Reasoning Models (LRMs) achieve strong performance in mathematics, code generation, and task planning, but their reliance on long chains of verbose "thinking" tokens leads to high latency, redundancy, and incoherent reasoning paths. Inspired by the Language of Thought Hypothesis, which posits that human reasoning operates over a symbolic, compositional mental language called Mentalese, we introduce a framework that trains models to reason in a similarly compact style. Mentalese encodes abstract reasoning as ultra-compressed, structured tokens, enabling models to solve complex problems with far fewer steps. To improve both efficiency and accuracy, we propose SHORTER LENGTH PREFERENCE OPTIMIZATION (SLPO), a reinforcement learning method that rewards concise solutions that stay correct, while still allowing longer reasoning when needed. Applied to Mentalese-aligned models, SLPO yields significantly higher compression rates by enabling concise reasoning that preserves the benefits of detailed thinking without the computational overhead. Across benchmarks including AIME 2024 and 2025, MinervaMath, OlympiadBench, Math500, and AMC, our ORION models produce reasoning traces with 4-16x fewer tokens, achieve up to 5x lower inference latency, and reduce training costs by 7-9x relative to the DeepSeek R1 Distilled model, while maintaining 90-98% of its accuracy. ORION also surpasses Claude and ChatGPT-4o by up to 5% in accuracy while maintaining 2x compression. These results show that Mentalese-style compressed reasoning offers a step toward human-like cognitive efficiency, enabling real-time, cost-effective reasoning without sacrificing accuracy.</p> + https://arxiv.org/abs/2511.22891 + Fri, 28 Nov 2025 05:41:55 +0000 - Downscaling Intelligence: Exploring Perception and Reasoning Bottlenecks in Small Multimodal Models - https://arxiv.org/abs/2511.17487 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.17487.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Mark Endo, Serena Yeung-Levy</p><p><b>Upvotes:</b> 9</p><p><b>Summary:</b> Scaling up multimodal models has enabled remarkable advances in visual understanding and reasoning, but practical demands call for smaller, efficient systems. In this work, we conduct a principled analysis of downscaling intelligence in multimodal models, examining how reduced large language model (LLM) capacity affects multimodal capabilities. Our initial findings reveal an interesting trend: LLM downscaling disproportionately affects visual capabilities, rather than abilities inherited from the LLM. We then examine whether this drop mainly reflects the expected decline in visual reasoning or a more fundamental loss of perceptual abilities. Isolating the effect of LLM downscaling on perception, we find performance still drops sharply, often matching or exceeding the impact on reasoning. To address this bottleneck, we introduce visual extraction tuning, which explicitly trains the model to extract instruction-relevant visual details consistently across tasks. With these extracted visual details, we then apply step-by-step reasoning to generate answers. Together, these components form our Extract+Think approach, setting a new standard for efficiency and performance in this space.</p> - https://arxiv.org/abs/2511.17487 - Fri, 21 Nov 2025 18:43:01 +0000 + MultiBanana: A Challenging Benchmark for Multi-Reference Text-to-Image Generation + https://arxiv.org/abs/2511.22989 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.22989.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Yuta Oshima, Daiki Miyake, Kohsei Matsutani, Yusuke Iwasawa, Masahiro Suzuki, Yutaka Matsuo, Hiroki Furuta</p><p><b>Upvotes:</b> 14</p><p><b>Summary:</b> Recent text-to-image generation models have acquired the ability of multi-reference generation and editing; the ability to inherit the appearance of subjects from multiple reference images and re-render them under new contexts. However, the existing benchmark datasets often focus on the generation with single or a few reference images, which prevents us from measuring the progress on how model performance advances or pointing out their weaknesses, under different multi-reference conditions. In addition, their task definitions are still vague, typically limited to axes such as "what to edit" or "how many references are given", and therefore fail to capture the intrinsic difficulty of multi-reference settings. To address this gap, we introduce MultiBanana, which is carefully designed to assesses the edge of model capabilities by widely covering multi-reference-specific problems at scale: (1) varying the number of references, (2) domain mismatch among references (e.g., photo vs. anime), (3) scale mismatch between reference and target scenes, (4) references containing rare concepts (e.g., a red banana), and (5) multilingual textual references for rendering. Our analysis among a variety of text-to-image models reveals their superior performances, typical failure modes, and areas for improvement. MultiBanana will be released as an open benchmark to push the boundaries and establish a standardized basis for fair comparison in multi-reference image generation. Our data and code are available at https://github.com/matsuolab/multibanana .</p> + https://arxiv.org/abs/2511.22989 + Fri, 28 Nov 2025 08:49:55 +0000 - Video-R4: Reinforcing Text-Rich Video Reasoning with Visual Rumination - https://arxiv.org/abs/2511.17490 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.17490.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Yolo Yunlong Tang, Daiki Shimada, Hang Hua, Chao Huang, Jing Bi, Rogerio Feris, Chenliang Xu</p><p><b>Upvotes:</b> 20</p><p><b>Summary:</b> Understanding text-rich videos requires reading small, transient textual cues that often demand repeated inspection. Yet most video QA models rely on single-pass perception over fixed frames, leading to hallucinations and failures on fine-grained evidence. Inspired by how humans pause, zoom, and re-read critical regions, we introduce Video-R4 (Reinforcing Text-Rich Video Reasoning with Visual Rumination), a video reasoning LMM that performs visual rumination: iteratively selecting frames, zooming into informative regions, re-encoding retrieved pixels, and updating its reasoning state. We construct two datasets with executable rumination trajectories: Video-R4-CoT-17k for supervised practice and Video-R4-RL-30k for reinforcement learning. We propose a multi-stage rumination learning framework that progressively finetunes a 7B LMM to learn atomic and mixing visual operations via SFT and GRPO-based RL. Video-R4-7B achieves state-of-the-art results on M4-ViteVQA and further generalizes to multi-page document QA, slides QA, and generic video QA, demonstrating that iterative rumination is an effective paradigm for pixel-grounded multimodal reasoning.</p> - https://arxiv.org/abs/2511.17490 - Fri, 21 Nov 2025 18:47:09 +0000 + DualCamCtrl: Dual-Branch Diffusion Model for Geometry-Aware Camera-Controlled Video Generation + https://arxiv.org/abs/2511.23127 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.23127.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Hongfei Zhang, Kanghao Chen, Zixin Zhang, Harold Haodong Chen, Yuanhuiyi Lyu, Yuqi Zhang, Shuai Yang, Kun Zhou, Yingcong Chen</p><p><b>Upvotes:</b> 41</p><p><b>Summary:</b> This paper presents DualCamCtrl, a novel end-to-end diffusion model for camera-controlled video generation. Recent works have advanced this field by representing camera poses as ray-based conditions, yet they often lack sufficient scene understanding and geometric awareness. DualCamCtrl specifically targets this limitation by introducing a dual-branch framework that mutually generates camera-consistent RGB and depth sequences. To harmonize these two modalities, we further propose the Semantic Guided Mutual Alignment (SIGMA) mechanism, which performs RGB-depth fusion in a semantics-guided and mutually reinforced manner. These designs collectively enable DualCamCtrl to better disentangle appearance and geometry modeling, generating videos that more faithfully adhere to the specified camera trajectories. Additionally, we analyze and reveal the distinct influence of depth and camera poses across denoising stages and further demonstrate that early and late stages play complementary roles in forming global structure and refining local details. Extensive experiments demonstrate that DualCamCtrl achieves more consistent camera-controlled video generation, with over 40\% reduction in camera motion errors compared with prior methods. Our project page: https://soyouthinkyoucantell.github.io/dualcamctrl-page/</p> + https://arxiv.org/abs/2511.23127 + Fri, 28 Nov 2025 12:19:57 +0000 - RynnVLA-002: A Unified Vision-Language-Action and World Model - https://arxiv.org/abs/2511.17502 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.17502.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Jun Cen, Siteng Huang, Yuqian Yuan, Hangjie Yuan, Chaohui Yu, Yuming Jiang, Jiayan Guo, Kehan Li, Hao Luo, Fan Wang, Xin Li, Deli Zhao, Hao Chen</p><p><b>Upvotes:</b> 24</p><p><b>Summary:</b> We introduce RynnVLA-002, a unified Vision-Language-Action (VLA) and world model. The world model leverages action and visual inputs to predict future image states, learning the underlying physics of the environment to refine action generation. Conversely, the VLA model produces subsequent actions from image observations, enhancing visual understanding and supporting the world model's image generation. The unified framework of RynnVLA-002 enables joint learning of environmental dynamics and action planning. Our experiments show that RynnVLA-002 surpasses individual VLA and world models, demonstrating their mutual enhancement. We evaluate RynnVLA-002 in both simulation and real-world robot tasks. RynnVLA-002 achieves 97.4% success rate on the LIBERO simulation benchmark without pretraining, while in real-world LeRobot experiments, its integrated world model boosts the overall success rate by 50%.</p> - https://arxiv.org/abs/2511.17502 - Fri, 21 Nov 2025 18:59:32 +0000 + Vision Bridge Transformer at Scale + https://arxiv.org/abs/2511.23199 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.23199.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Zhenxiong Tan, Zeqing Wang, Xingyi Yang, Songhua Liu, Xinchao Wang</p><p><b>Upvotes:</b> 41</p><p><b>Summary:</b> We introduce Vision Bridge Transformer (ViBT), a large-scale instantiation of Brownian Bridge Models designed for conditional generation. Unlike traditional diffusion models that transform noise into data, Bridge Models directly model the trajectory between inputs and outputs, creating an efficient data-to-data translation paradigm. By scaling these models to 20B and 1.3B parameters, we demonstrate their effectiveness for image and video translation tasks. To support this scale, we adopt a Transformer architecture and propose a variance-stabilized velocity-matching objective for robust training. Together, these advances highlight the power of scaling Bridge Models for instruction-based image editing and complex video translation.</p> + https://arxiv.org/abs/2511.23199 + Fri, 28 Nov 2025 14:03:39 +0000 - M3-Bench: Multi-Modal, Multi-Hop, Multi-Threaded Tool-Using MLLM Agent Benchmark - https://arxiv.org/abs/2511.17729 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.17729.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Yang Zhou, Mingyu Zhao, Zhenting Wang, Difei Gu, Bangwei Guo, Ruosong Ye, Ligong Han, Can Jin, Dimitris N. Metaxas</p><p><b>Upvotes:</b> 16</p><p><b>Summary:</b> We present M^3-Bench, the first benchmark for evaluating multimodal tool use under the Model Context Protocol. The benchmark targets realistic, multi-hop and multi-threaded workflows that require visual grounding and textual reasoning, cross-tool dependencies, and persistence of intermediate resources across steps. We introduce a similarity-driven alignment that serializes each tool call, embeds signatures with a sentence encoder, and performs similarity-bucketed Hungarian matching to obtain auditable one-to-one correspondences. On top of this alignment, we report interpretable metrics that decouple semantic fidelity from workflow consistency. The benchmark spans 28 servers with 231 tools, and provides standardized trajectories curated through an Executor & Judge pipeline with human verification; an auxiliary four large language models (LLMs) judge ensemble reports end-task Task Completion and information grounding. Evaluations of representative state-of-the-art Multimodal LLMs (MLLMs) reveal persistent gaps in multimodal MCP tool use, particularly in argument fidelity and structure consistency, underscoring the need for methods that jointly reason over images, text, and tool graphs. Our Benchmark's anonymous repository is at https://github.com/EtaYang10th/Open-M3-Bench</p> - https://arxiv.org/abs/2511.17729 - Fri, 21 Nov 2025 19:27:02 +0000 + Every Token Counts: Generalizing 16M Ultra-Long Context in Large Language Models + https://arxiv.org/abs/2511.23319 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.23319.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Xiang Hu, Zhanchao Zhou, Ruiqi Liang, Zehuan Li, Wei Wu, Jianguo Li</p><p><b>Upvotes:</b> 21</p><p><b>Summary:</b> This work explores the challenge of building ``Machines that Can Remember'', framing long-term memory as the problem of efficient ultra-long context modeling. We argue that this requires three key properties: sparsity, random-access flexibility, and length generalization. To address ultra-long-context modeling, we leverage Hierarchical Sparse Attention (HSA), a novel attention mechanism that satisfies all three properties. We integrate HSA into Transformers to build HSA-UltraLong, which is an 8B-parameter MoE model trained on over 8 trillion tokens and is rigorously evaluated on different tasks with in-domain and out-of-domain context lengths to demonstrate its capability in handling ultra-long contexts. Results show that our model performs comparably to full-attention baselines on in-domain lengths while achieving over 90\% accuracy on most in-context retrieval tasks with contexts up to 16M. This report outlines our experimental insights and open problems, contributing a foundation for future research in ultra-long context modeling.</p> + https://arxiv.org/abs/2511.23319 + Fri, 28 Nov 2025 16:17:53 +0000 - Target-Bench: Can World Models Achieve Mapless Path Planning with Semantic Targets? - https://arxiv.org/abs/2511.17792 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.17792.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Dingrui Wang, Hongyuan Ye, Zhihao Liang, Zhexiao Sun, Zhaowei Lu, Yuchen Zhang, Yuyu Zhao, Yuan Gao, Marvin Seegert, Finn Schäfer, Haotong Qin, Wei Li, Luigi Palmieri, Felix Jahncke, Mattia Piccinini, Johannes Betz</p><p><b>Upvotes:</b> 3</p><p><b>Summary:</b> While recent world models generate highly realistic videos, their ability to perform robot path planning remains unclear and unquantified. We introduce Target-Bench, the first benchmark specifically designed to evaluate world models on mapless path planning toward semantic targets in real-world environments. Target-Bench provides 450 robot-collected video sequences spanning 45 semantic categories with SLAM-based ground truth trajectories. Our evaluation pipeline recovers camera motion from generated videos and measures planning performance using five complementary metrics that quantify target-reaching capability, trajectory accuracy, and directional consistency. We evaluate state-of-the-art models including Sora 2, Veo 3.1, and the Wan series. The best off-the-shelf model (Wan2.2-Flash) achieves only 0.299 overall score, revealing significant limitations in current world models for robotic planning tasks. We show that fine-tuning an open-source 5B-parameter model on only 325 scenarios from our dataset achieves 0.345 overall score -- an improvement of more than 400% over its base version (0.066) and 15% higher than the best off-the-shelf model. We will open-source the code and dataset.</p> - https://arxiv.org/abs/2511.17792 - Fri, 21 Nov 2025 21:36:02 +0000 + Flow Straighter and Faster: Efficient One-Step Generative Modeling via MeanFlow on Rectified Trajectories + https://arxiv.org/abs/2511.23342 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.23342.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Xinxi Zhang, Shiwei Tan, Quang Nguyen, Quan Dao, Ligong Han, Xiaoxiao He, Tunyu Zhang, Alen Mrdovic, Dimitris Metaxas</p><p><b>Upvotes:</b> 14</p><p><b>Summary:</b> Flow-based generative models have recently demonstrated strong performance, yet sampling typically relies on expensive numerical integration of ordinary differential equations (ODEs). Rectified Flow enables one-step sampling by learning nearly straight probability paths, but achieving such straightness requires multiple computationally intensive reflow iterations. MeanFlow achieves one-step generation by directly modeling the average velocity over time; however, when trained on highly curved flows, it suffers from slow convergence and noisy supervision. To address these limitations, we propose Rectified MeanFlow, a framework that models the mean velocity field along the rectified trajectory using only a single reflow step. This eliminates the need for perfectly straightened trajectories while enabling efficient training. Furthermore, we introduce a simple yet effective truncation heuristic that aims to reduce residual curvature and further improve performance. Extensive experiments on ImageNet at 64, 256, and 512 resolutions show that Re-MeanFlow consistently outperforms prior one-step flow distillation and Rectified Flow methods in both sample quality and training efficiency. Code is available at https://github.com/Xinxi-Zhang/Re-MeanFlow.</p> + https://arxiv.org/abs/2511.23342 + Fri, 28 Nov 2025 16:50:08 +0000 - Pillar-0: A New Frontier for Radiology Foundation Models - https://arxiv.org/abs/2511.17803 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.17803.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Kumar Krishna Agrawal, Longchao Liu, Long Lian, Michael Nercessian, Natalia Harguindeguy, Yufu Wu, Peter Mikhael, Gigin Lin, Lecia V. Sequist, Florian Fintelmann, Trevor Darrell, Yutong Bai, Maggie Chung, Adam Yala</p><p><b>Upvotes:</b> 18</p><p><b>Summary:</b> Radiology plays an integral role in modern medicine, yet rising imaging volumes have far outpaced workforce growth. Foundation models offer a path toward assisting with the full spectrum of radiology tasks, but existing medical models remain limited: they process volumetric CT and MRI as low-fidelity 2D slices, discard critical grayscale contrast information, and lack evaluation frameworks that reflect real clinical practice. We introduce Pillar-0, a radiology foundation model pretrained on 42,990 abdomen-pelvis CTs, 86,411 chest CTs, 14,348 head CTs, and 11,543 breast MRIs from a large academic center, together with RATE, a scalable framework that extracts structured labels for 366 radiologic findings with near-perfect accuracy using LLMs. Across internal test sets of 14,230 abdomen-pelvis CTs, 10,646 chest CTs, 4,906 head CTs, and 1,585 breast MRIs, Pillar-0 establishes a new performance frontier, achieving mean AUROCs of 86.4, 88.0, 90.1, and 82.9, outperforming MedGemma (Google), MedImageInsight (Microsoft), Lingshu (Alibaba), and Merlin (Stanford) by 7.8-15.8 AUROC points and ranking best in 87.2\% (319/366) tasks. Pillar-0 similarly outperforms all baselines in an external validation on the Stanford Abdominal CT dataset, including Merlin (82.2 vs 80.6 AUROC). Pillar-0 extends to tasks beyond its pretraining, such as long-horizon lung cancer risk prediction, where it improves upon the state-of-the-art Sybil by 3.0 C-index points on NLST, and generalizes with gains of 5.9 (MGH) and 1.9 (CGMH). In brain hemorrhage detection, Pillar-0 obtained a >95 AUROC when using only 1/20th of the data of the next most sample efficient baseline. Pillar-0 and RATE together provide an open, clinically rigorous foundation for building high-performance radiology systems, enabling applications that were previously infeasible due to computational, data, and evaluation constraints.</p> - https://arxiv.org/abs/2511.17803 - Fri, 21 Nov 2025 21:50:34 +0000 + SimScale: Learning to Drive via Real-World Simulation at Scale + https://arxiv.org/abs/2511.23369 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.23369.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Haochen Tian, Tianyu Li, Haochen Liu, Jiazhi Yang, Yihang Qiu, Guang Li, Junli Wang, Yinfeng Gao, Zhang Zhang, Liang Wang, Hangjun Ye, Tieniu Tan, Long Chen, Hongyang Li</p><p><b>Upvotes:</b> 37</p><p><b>Summary:</b> Achieving fully autonomous driving systems requires learning rational decisions in a wide span of scenarios, including safety-critical and out-of-distribution ones. However, such cases are underrepresented in real-world corpus collected by human experts. To complement for the lack of data diversity, we introduce a novel and scalable simulation framework capable of synthesizing massive unseen states upon existing driving logs. Our pipeline utilizes advanced neural rendering with a reactive environment to generate high-fidelity multi-view observations controlled by the perturbed ego trajectory. Furthermore, we develop a pseudo-expert trajectory generation mechanism for these newly simulated states to provide action supervision. Upon the synthesized data, we find that a simple co-training strategy on both real-world and simulated samples can lead to significant improvements in both robustness and generalization for various planning methods on challenging real-world benchmarks, up to +6.8 EPDMS on navhard and +2.9 on navtest. More importantly, such policy improvement scales smoothly by increasing simulation data only, even without extra real-world data streaming in. We further reveal several crucial findings of such a sim-real learning system, which we term SimScale, including the design of pseudo-experts and the scaling properties for different policy architectures. Our simulation data and code would be released.</p> + https://arxiv.org/abs/2511.23369 + Fri, 28 Nov 2025 17:17:38 +0000 - MobileVLA-R1: Reinforcing Vision-Language-Action for Mobile Robots - https://arxiv.org/abs/2511.17889 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.17889.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Ting Huang, Dongjian Li, Rui Yang, Zeyu Zhang, Zida Yang, Hao Tang</p><p><b>Upvotes:</b> 5</p><p><b>Summary:</b> Grounding natural-language instructions into continuous control for quadruped robots remains a fundamental challenge in vision language action. Existing methods struggle to bridge high-level semantic reasoning and low-level actuation, leading to unstable grounding and weak generalization in the real world. To address these issues, we present MobileVLA-R1, a unified vision-language-action framework that enables explicit reasoning and continuous control for quadruped robots. We construct MobileVLA-CoT, a large-scale dataset of multi-granularity chain-of-thought (CoT) for embodied trajectories, providing structured reasoning supervision for alignment. Built upon this foundation, we introduce a two-stage training paradigm that combines supervised CoT alignment with GRPO reinforcement learning to enhance reasoning consistency, control stability, and long-horizon execution. Extensive evaluations on VLN and VLA tasks demonstrate superior performance over strong baselines, with approximately a 5% improvement. Real-world deployment on a quadruped robot validates robust performance in complex environments. Code: https://github.com/AIGeeksGroup/MobileVLA-R1. Website: https://aigeeksgroup.github.io/MobileVLA-R1.</p> - https://arxiv.org/abs/2511.17889 - Sat, 22 Nov 2025 02:34:10 +0000 + LFM2 Technical Report + https://arxiv.org/abs/2511.23404 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.23404.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Alexander Amini, Anna Banaszak, Harold Benoit, Arthur Böök, Tarek Dakhran, Song Duong, Alfred Eng, Fernando Fernandes, Marc Härkönen, Anne Harrington, Ramin Hasani, Saniya Karwa, Yuri Khrustalev, Maxime Labonne, Mathias Lechner, Valentine Lechner, Simon Lee, Zetian Li, Noel Loo, Jacob Marks, Edoardo Mosca, Samuel J. Paech, Paul Pak, Rom N. Parnichkun, Alex Quach, Ryan Rogers, Daniela Rus, Nayan Saxena, Bettina Schlager, Tim Seyde, Jimmy T. H. Smith, Aditya Tadimeti, Neehal Tumma</p><p><b>Upvotes:</b> 34</p><p><b>Summary:</b> We present LFM2, a family of Liquid Foundation Models designed for efficient on-device deployment and strong task capabilities. Using hardware-in-the-loop architecture search under edge latency and memory constraints, we obtain a compact hybrid backbone that combines gated short convolutions with a small number of grouped query attention blocks, delivering up to 2x faster prefill and decode on CPUs compared to similarly sized models. The LFM2 family covers 350M-8.3B parameters, including dense models (350M, 700M, 1.2B, 2.6B) and a mixture-of-experts variant (8.3B total, 1.5B active), all with 32K context length. LFM2's training pipeline includes a tempered, decoupled Top-K knowledge distillation objective that avoids support mismatch; curriculum learning with difficulty-ordered data; and a three-stage post-training recipe of supervised fine-tuning, length-normalized preference optimization, and model merging. Pre-trained on 10-12T tokens, LFM2 models achieve strong results across diverse benchmarks; for example, LFM2-2.6B reaches 79.56% on IFEval and 82.41% on GSM8K. We further build multimodal and retrieval variants: LFM2-VL for vision-language tasks, LFM2-Audio for speech, and LFM2-ColBERT for retrieval. LFM2-VL supports tunable accuracy-latency tradeoffs via token-efficient visual processing, while LFM2-Audio separates audio input and output pathways to enable real-time speech-to-speech interaction competitive with models 3x larger. LFM2-ColBERT provides a low-latency encoder for queries and documents, enabling high-performance retrieval across multiple languages. All models are released with open weights and deployment packages for ExecuTorch, llama.cpp, and vLLM, making LFM2 a practical base for edge applications that need fast, memory-efficient inference and strong task capabilities.</p> + https://arxiv.org/abs/2511.23404 + Fri, 28 Nov 2025 17:56:35 +0000 - SciEducator: Scientific Video Understanding and Educating via Deming-Cycle Multi-Agent System - https://arxiv.org/abs/2511.17943 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.17943.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Zhiyu Xu, Weilong Yan, Yufei Shi, Xin Meng, Tao He, Huiping Zhuang, Ming Li, Hehe Fan</p><p><b>Upvotes:</b> 2</p><p><b>Summary:</b> Recent advancements in multimodal large language models (MLLMs) and video agent systems have significantly improved general video understanding. However, when applied to scientific video understanding and educating, a domain that demands external professional knowledge integration and rigorous step-wise reasoning, existing approaches often struggle. To bridge this gap, we propose SciEducator, the first iterative self-evolving multi-agent system for scientific video comprehension and education. Rooted in the classical Deming Cycle from management science, our design reformulates its Plan-Do-Study-Act philosophy into a self-evolving reasoning and feedback mechanism, which facilitates the interpretation of intricate scientific activities in videos. Moreover, SciEducator can produce multimodal educational content tailored to specific scientific processes, including textual instructions, visual guides, audio narrations, and interactive references. To support evaluation, we construct SciVBench, a benchmark consisting of 500 expert-verified and literature-grounded science QA pairs across five categories, covering physical, chemical, and everyday phenomena. Extensive experiments demonstrate that SciEducator substantially outperforms leading closed-source MLLMs (e.g., Gemini, GPT-4o) and state-of-the-art video agents on the benchmark, establishing a new paradigm for the community.</p> - https://arxiv.org/abs/2511.17943 - Sat, 22 Nov 2025 06:54:16 +0000 + AnyTalker: Scaling Multi-Person Talking Video Generation with Interactivity Refinement + https://arxiv.org/abs/2511.23475 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.23475.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Zhizhou Zhong, Yicheng Ji, Zhe Kong, Yiying Liu, Jiarui Wang, Jiasun Feng, Lupeng Liu, Xiangyi Wang, Yanjia Li, Yuqing She, Ying Qin, Huan Li, Shuiyang Mao, Wei Liu, Wenhan Luo</p><p><b>Upvotes:</b> 41</p><p><b>Summary:</b> Recently, multi-person video generation has started to gain prominence. While a few preliminary works have explored audio-driven multi-person talking video generation, they often face challenges due to the high costs of diverse multi-person data collection and the difficulty of driving multiple identities with coherent interactivity. To address these challenges, we propose AnyTalker, a multi-person generation framework that features an extensible multi-stream processing architecture. Specifically, we extend Diffusion Transformer's attention block with a novel identity-aware attention mechanism that iteratively processes identity-audio pairs, allowing arbitrary scaling of drivable identities. Besides, training multi-person generative models demands massive multi-person data. Our proposed training pipeline depends solely on single-person videos to learn multi-person speaking patterns and refines interactivity with only a few real multi-person clips. Furthermore, we contribute a targeted metric and dataset designed to evaluate the naturalness and interactivity of the generated multi-person videos. Extensive experiments demonstrate that AnyTalker achieves remarkable lip synchronization, visual quality, and natural interactivity, striking a favorable balance between data costs and identity scalability.</p> + https://arxiv.org/abs/2511.23475 + Fri, 28 Nov 2025 18:59:01 +0000 - Plan-X: Instruct Video Generation via Semantic Planning - https://arxiv.org/abs/2511.17986 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.17986.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Lun Huang, You Xie, Hongyi Xu, Tianpei Gu, Chenxu Zhang, Guoxian Song, Zenan Li, Xiaochen Zhao, Linjie Luo, Guillermo Sapiro</p><p><b>Upvotes:</b> 16</p><p><b>Summary:</b> Diffusion Transformers have demonstrated remarkable capabilities in visual synthesis, yet they often struggle with high-level semantic reasoning and long-horizon planning. This limitation frequently leads to visual hallucinations and mis-alignments with user instructions, especially in scenarios involving complex scene understanding, human-object interactions, multi-stage actions, and in-context motion reasoning. To address these challenges, we propose Plan-X, a framework that explicitly enforces high-level semantic planning to instruct video generation process. At its core lies a Semantic Planner, a learnable multimodal language model that reasons over the user's intent from both text prompts and visual context, and autoregressively generates a sequence of text-grounded spatio-temporal semantic tokens. These semantic tokens, complementary to high-level text prompt guidance, serve as structured "semantic sketches" over time for the video diffusion model, which has its strength at synthesizing high-fidelity visual details. Plan-X effectively integrates the strength of language models in multimodal in-context reasoning and planning, together with the strength of diffusion models in photorealistic video synthesis. Extensive experiments demonstrate that our framework substantially reduces visual hallucinations and enables fine-grained, instruction-aligned video generation consistent with multimodal context.</p> - https://arxiv.org/abs/2511.17986 - Sat, 22 Nov 2025 08:59:09 +0000 + What about gravity in video generation? Post-Training Newton's Laws with Verifiable Rewards + https://arxiv.org/abs/2512.00425 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2512.00425.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Minh-Quan Le, Yuanzhi Zhu, Vicky Kalogeiton, Dimitris Samaras</p><p><b>Upvotes:</b> 45</p><p><b>Summary:</b> 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 NewtonRewards, the first physics-grounded post-training framework for video generation based on verifiable rewards. Instead of relying on human or VLM feedback, NewtonRewards extracts 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 NewtonRewards on five Newtonian Motion Primitives (free fall, horizontal/parabolic throw, and ramp sliding down/up) using our newly constructed large-scale benchmark, NewtonBench-60K. Across all primitives in visual and physics metrics, 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.</p> + https://arxiv.org/abs/2512.00425 + Sat, 29 Nov 2025 10:04:50 +0000 - Extracting Interaction-Aware Monosemantic Concepts in Recommender Systems - https://arxiv.org/abs/2511.18024 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.18024.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Dor Arviv, Yehonatan Elisha, Oren Barkan, Noam Koenigstein</p><p><b>Upvotes:</b> 1</p><p><b>Summary:</b> We present a method for extracting monosemantic neurons, defined as latent dimensions that align with coherent and interpretable concepts, from user and item embeddings in recommender systems. Our approach employs a Sparse Autoencoder (SAE) to reveal semantic structure within pretrained representations. In contrast to work on language models, monosemanticity in recommendation must preserve the interactions between separate user and item embeddings. To achieve this, we introduce a prediction aware training objective that backpropagates through a frozen recommender and aligns the learned latent structure with the model's user-item affinity predictions. The resulting neurons capture properties such as genre, popularity, and temporal trends, and support post hoc control operations including targeted filtering and content promotion without modifying the base model. Our method generalizes across different recommendation models and datasets, providing a practical tool for interpretable and controllable personalization. Code and evaluation resources are available at https://github.com/DeltaLabTLV/Monosemanticity4Rec.</p> - https://arxiv.org/abs/2511.18024 - Sat, 22 Nov 2025 11:27:32 +0000 + SCALE: Selective Resource Allocation for Overcoming Performance Bottlenecks in Mathematical Test-time Scaling + https://arxiv.org/abs/2512.00466 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2512.00466.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Yang Xiao, Chunpu Xu, Ruifeng Yuan, Jiashuo Wang, Wenjie Li, Pengfei Liu</p><p><b>Upvotes:</b> 8</p><p><b>Summary:</b> Test-time compute scaling has emerged as a powerful paradigm for enhancing mathematical reasoning in large language models (LLMs) by allocating additional computational resources during inference. However, current methods employ uniform resource distribution across all reasoning sub-problems, creating fundamental bottlenecks where challenging sub-problems receive insufficient attention while routine operations consume disproportionate resources. This uniform allocation creates performance bottlenecks where additional computational resources yield diminishing returns. Inspired by dual-process theory, we propose SCALE (Selective Resource Allocation), a framework that selectively allocates computational resources based on sub-problem difficulty. SCALE operates through four stages: (1) problem decomposition into sequential reasoning sub-problems, (2) difficulty assessment of each sub-problem to distinguish between routine operations and computationally challenging sub-problems, (3) selective processing mode assignment between System 1 for simple sub-problems and System 2 for complex ones, and (4) sequential execution with context propagation. By concentrating resources on challenging sub-problems while processing routine operations efficiently, SCALE achieves substantial performance improvements with superior resource utilization. Extensive experiments demonstrate that SCALE significantly outperforms uniform scaling baselines, achieving accuracy improvements of up to 13.75 percentage points (57.50% to 71.25% on AIME25) while reducing computational costs by 33%-53%, representing a major advance in test-time scaling that addresses fundamental limitations of current approaches.</p> + https://arxiv.org/abs/2512.00466 + Sat, 29 Nov 2025 12:38:07 +0000 - Fidelity-Aware Recommendation Explanations via Stochastic Path Integration - https://arxiv.org/abs/2511.18047 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.18047.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Oren Barkan, Yahlly Schein, Yehonatan Elisha, Veronika Bogina, Mikhail Baklanov, Noam Koenigstein</p><p><b>Upvotes:</b> 1</p><p><b>Summary:</b> Explanation fidelity, which measures how accurately an explanation reflects a model's true reasoning, remains critically underexplored in recommender systems. We introduce SPINRec (Stochastic Path Integration for Neural Recommender Explanations), a model-agnostic approach that adapts path-integration techniques to the sparse and implicit nature of recommendation data. To overcome the limitations of prior methods, SPINRec employs stochastic baseline sampling: instead of integrating from a fixed or unrealistic baseline, it samples multiple plausible user profiles from the empirical data distribution and selects the most faithful attribution path. This design captures the influence of both observed and unobserved interactions, yielding more stable and personalized explanations. We conduct the most comprehensive fidelity evaluation to date across three models (MF, VAE, NCF), three datasets (ML1M, Yahoo! Music, Pinterest), and a suite of counterfactual metrics, including AUC-based perturbation curves and fixed-length diagnostics. SPINRec consistently outperforms all baselines, establishing a new benchmark for faithful explainability in recommendation. Code and evaluation tools are publicly available at https://github.com/DeltaLabTLV/SPINRec.</p> - https://arxiv.org/abs/2511.18047 - Sat, 22 Nov 2025 12:59:04 +0000 + Wikontic: Constructing Wikidata-Aligned, Ontology-Aware Knowledge Graphs with Large Language Models + https://arxiv.org/abs/2512.00590 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2512.00590.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Alla Chepurova, Aydar Bulatov, Yuri Kuratov, Mikhail Burtsev</p><p><b>Upvotes:</b> 34</p><p><b>Summary:</b> Knowledge graphs (KGs) provide structured, verifiable grounding for large language models (LLMs), but current LLM-based systems commonly use KGs as auxiliary structures for text retrieval, leaving their intrinsic quality underexplored. In this work, we propose Wikontic, a multi-stage pipeline that constructs KGs from open-domain text by extracting candidate triplets with qualifiers, enforcing Wikidata-based type and relation constraints, and normalizing entities to reduce duplication. The resulting KGs are compact, ontology-consistent, and well-connected; on MuSiQue, the correct answer entity appears in 96% of generated triplets. On HotpotQA, our triplets-only setup achieves 76.0 F1, and on MuSiQue 59.8 F1, matching or surpassing several retrieval-augmented generation baselines that still require textual context. In addition, Wikontic attains state-of-the-art information-retention performance on the MINE-1 benchmark (86%), outperforming prior KG construction methods. Wikontic is also efficient at build time: KG construction uses less than 1,000 output tokens, about 3times fewer than AriGraph and <1/20 of GraphRAG. The proposed pipeline enhances the quality of the generated KG and offers a scalable solution for leveraging structured knowledge in LLMs.</p> + https://arxiv.org/abs/2512.00590 + Sat, 29 Nov 2025 18:44:25 +0000 - UltraFlux: Data-Model Co-Design for High-quality Native 4K Text-to-Image Generation across Diverse Aspect Ratios - https://arxiv.org/abs/2511.18050 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.18050.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Tian Ye, Song Fei, Lei Zhu</p><p><b>Upvotes:</b> 36</p><p><b>Summary:</b> Diffusion transformers have recently delivered strong text-to-image generation around 1K resolution, but we show that extending them to native 4K across diverse aspect ratios exposes a tightly coupled failure mode spanning positional encoding, VAE compression, and optimization. Tackling any of these factors in isolation leaves substantial quality on the table. We therefore take a data-model co-design view and introduce UltraFlux, a Flux-based DiT trained natively at 4K on MultiAspect-4K-1M, a 1M-image 4K corpus with controlled multi-AR coverage, bilingual captions, and rich VLM/IQA metadata for resolution- and AR-aware sampling. On the model side, UltraFlux couples (i) Resonance 2D RoPE with YaRN for training-window-, frequency-, and AR-aware positional encoding at 4K; (ii) a simple, non-adversarial VAE post-training scheme that improves 4K reconstruction fidelity; (iii) an SNR-Aware Huber Wavelet objective that rebalances gradients across timesteps and frequency bands; and (iv) a Stage-wise Aesthetic Curriculum Learning strategy that concentrates high-aesthetic supervision on high-noise steps governed by the model prior. Together, these components yield a stable, detail-preserving 4K DiT that generalizes across wide, square, and tall ARs. On the Aesthetic-Eval at 4096 benchmark and multi-AR 4K settings, UltraFlux consistently outperforms strong open-source baselines across fidelity, aesthetic, and alignment metrics, and-with a LLM prompt refiner-matches or surpasses the proprietary Seedream 4.0.</p> - https://arxiv.org/abs/2511.18050 - Sat, 22 Nov 2025 13:07:21 +0000 + SpeContext: Enabling Efficient Long-context Reasoning with Speculative Context Sparsity in LLMs + https://arxiv.org/abs/2512.00722 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2512.00722.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Jiaming Xu, Jiayi Pan, Hanzhen Wang, Yongkang Zhou, Jiancai Ye, Yu Wang, Guohao Dai</p><p><b>Upvotes:</b> 14</p><p><b>Summary:</b> 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.</p> + https://arxiv.org/abs/2512.00722 + Sun, 30 Nov 2025 04:32:43 +0000 - MASS: Motion-Aware Spatial-Temporal Grounding for Physics Reasoning and Comprehension in Vision-Language Models - https://arxiv.org/abs/2511.18373 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.18373.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Xiyang Wu, Zongxia Li, Jihui Jin, Guangyao Shi, Gouthaman KV, Vishnu Raj, Nilotpal Sinha, Jingxi Chen, Fan Du, Dinesh Manocha</p><p><b>Upvotes:</b> 5</p><p><b>Summary:</b> Vision Language Models (VLMs) perform well on standard video tasks but struggle with physics-driven reasoning involving motion dynamics and spatial interactions. This limitation reduces their ability to interpret real or AI-generated content (AIGC) videos and to generate physically consistent content. We present an approach that addresses this gap by translating physical-world context cues into interpretable representations aligned with VLMs' perception, comprehension, and reasoning. We introduce MASS-Bench, a comprehensive benchmark consisting of 4,350 real-world and AIGC videos and 8,361 free-form video question-answering pairs focused on physics-related comprehension tasks, with detailed annotations including visual detections, sub-segment grounding, and full-sequence 3D motion tracking of entities. We further present MASS, a model-agnostic method that injects spatial-temporal signals into the VLM language space via depth-based 3D encoding and visual grounding, coupled with a motion tracker for object dynamics. To strengthen cross-modal alignment and reasoning, we apply reinforcement fine-tuning. Experiments and ablations show that our refined VLMs outperform comparable and larger baselines, as well as prior state-of-the-art models, by 8.7% and 6.0%, achieving performance comparable to close-source SoTA VLMs such as Gemini-2.5-Flash on physics reasoning and comprehension. These results validate the effectiveness of our approach.</p> - https://arxiv.org/abs/2511.18373 - Sun, 23 Nov 2025 09:43:44 +0000 + Accelerating Streaming Video Large Language Models via Hierarchical Token Compression + https://arxiv.org/abs/2512.00891 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2512.00891.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Yiyu Wang, Xuyang Liu, Xiyan Gui, Xinying Lin, Boxue Yang, Chenfei Liao, Tailai Chen, Linfeng Zhang</p><p><b>Upvotes:</b> 14</p><p><b>Summary:</b> 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 Streaming Token Compression (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: STC-Cacher, which reduces ViT encoding overhead by caching and reusing features from temporally similar frames, and 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 99\% of accuracy on the ReKV framework while reducing ViT encoding latency and LLM pre-filling latency by 24.5\% and 45.3\%.</p> + https://arxiv.org/abs/2512.00891 + Sun, 30 Nov 2025 13:44:28 +0000 - Future Is Unevenly Distributed: Forecasting Ability of LLMs Depends on What We're Asking - https://arxiv.org/abs/2511.18394 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.18394.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Chinmay Karkar, Paras Chopra</p><p><b>Upvotes:</b> 1</p><p><b>Summary:</b> Large Language Models (LLMs) demonstrate partial forecasting competence across social, political, and economic events. Yet, their predictive ability varies sharply with domain structure and prompt framing. We investigate how forecasting performance varies with different model families on real-world questions about events that happened beyond the model cutoff date. We analyze how context, question type, and external knowledge affect accuracy and calibration, and how adding factual news context modifies belief formation and failure modes. Our results show that forecasting ability is highly variable as it depends on what, and how, we ask.</p> - https://arxiv.org/abs/2511.18394 - Sun, 23 Nov 2025 10:41:19 +0000 + WUSH: Near-Optimal Adaptive Transforms for LLM Quantization + https://arxiv.org/abs/2512.00956 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2512.00956.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Jiale Chen, Vage Egiazarian, Torsten Hoefler, Dan Alistarh</p><p><b>Upvotes:</b> 17</p><p><b>Summary:</b> 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.</p> + https://arxiv.org/abs/2512.00956 + Sun, 30 Nov 2025 16:17:34 +0000 - General Agentic Memory Via Deep Research - https://arxiv.org/abs/2511.18423 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.18423.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> B. Y. Yan, Chaofan Li, Hongjin Qian, Shuqi Lu, Zheng Liu</p><p><b>Upvotes:</b> 146</p><p><b>Summary:</b> Memory is critical for AI agents, yet the widely-adopted static memory, aiming to create readily available memory in advance, is inevitably subject to severe information loss. To address this limitation, we propose a novel framework called general agentic memory (GAM). GAM follows the principle of "just-in time (JIT) compilation" where it focuses on creating optimized contexts for its client at runtime while keeping only simple but useful memory during the offline stage. To this end, GAM employs a duo-design with the following components. 1) Memorizer, which highlights key historical information using a lightweight memory, while maintaining complete historical information within a universal page-store. 2) Researcher, which retrieves and integrates useful information from the page-store for its online request guided by the pre-constructed memory. This design allows GAM to effectively leverage the agentic capabilities and test-time scalability of frontier large language models (LLMs), while also facilitating end-to-end performance optimization through reinforcement learning. In our experimental study, we demonstrate that GAM achieves substantial improvement on various memory-grounded task completion scenarios against existing memory systems.</p> - https://arxiv.org/abs/2511.18423 - Sun, 23 Nov 2025 12:29:33 +0000 + Lotus-2: Advancing Geometric Dense Prediction with Powerful Image Generative Model + https://arxiv.org/abs/2512.01030 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2512.01030.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Jing He, Haodong Li, Mingzhi Sheng, Ying-Cong Chen</p><p><b>Upvotes:</b> 16</p><p><b>Summary:</b> 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.</p> + https://arxiv.org/abs/2512.01030 + Sun, 30 Nov 2025 18:57:25 +0000 - Position: The Complexity of Perfect AI Alignment -- Formalizing the RLHF Trilemma - https://arxiv.org/abs/2511.19504 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.19504.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Subramanyam Sahoo, Aman Chadha, Vinija Jain, Divya Chaudhary</p><p><b>Upvotes:</b> 1</p><p><b>Summary:</b> Reinforcement Learning from Human Feedback (RLHF) is widely used for aligning large language models, yet practitioners face a persistent puzzle: improving safety often reduces fairness, scaling to diverse populations becomes computationally intractable, and making systems robust often amplifies majority biases. We formalize this tension as the Alignment Trilemma: no RLHF system can simultaneously achieve (i) epsilon-representativeness across diverse human values, (ii) polynomial tractability in sample and compute complexity, and (iii) delta-robustness against adversarial perturbations and distribution shift. Through a complexity-theoretic analysis integrating statistical learning theory and robust optimization, we prove that achieving both representativeness (epsilon <= 0.01) and robustness (delta <= 0.001) for global-scale populations requires Omega(2^{d_context}) operations, which is super-polynomial in the context dimensionality. We show that current RLHF implementations resolve this trilemma by sacrificing representativeness: they collect only 10^3--10^4 samples from homogeneous annotator pools while 10^7--10^8 samples are needed for true global representation. Our framework provides a unified explanation for documented RLHF pathologies including preference collapse, sycophancy, and systematic bias amplification. We conclude with concrete directions for navigating these fundamental trade-offs through strategic relaxations of alignment requirements.</p> - https://arxiv.org/abs/2511.19504 - Sun, 23 Nov 2025 20:23:23 +0000 + VLASH: Real-Time VLAs via Future-State-Aware Asynchronous Inference + https://arxiv.org/abs/2512.01031 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2512.01031.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Jiaming Tang, Yufei Sun, Yilong Zhao, Shang Yang, Yujun Lin, Zhuoyang Zhang, James Hou, Yao Lu, Zhijian Liu, Song Han</p><p><b>Upvotes:</b> 22</p><p><b>Summary:</b> 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</p> + https://arxiv.org/abs/2512.01031 + Sun, 30 Nov 2025 18:59:24 +0000 - CLaRa: Bridging Retrieval and Generation with Continuous Latent Reasoning - https://arxiv.org/abs/2511.18659 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.18659.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Jie He, Richard He Bai, Sinead Williamson, Jeff Z. Pan, Navdeep Jaitly, Yizhe Zhang</p><p><b>Upvotes:</b> 5</p><p><b>Summary:</b> Retrieval-augmented generation (RAG) enhances large language models (LLMs) with external knowledge but still suffers from long contexts and disjoint retrieval-generation optimization. In this work, we propose CLaRa (Continuous Latent Reasoning), a unified framework that performs embedding-based compression and joint optimization in a shared continuous space. To obtain semantically rich and retrievable compressed vectors, we introduce SCP, a key-preserving data synthesis framework using QA and paraphrase supervision. CLaRa then trains the reranker and generator end-to-end via a single language modeling loss, with gradients flowing through both modules using a differentiable top-k estimator. Theoretically, this unified optimization aligns retrieval relevance with answer quality. Experiments across multiple QA benchmarks show that CLaRa achieves state-of-the-art compression and reranking performance, often surpassing text-based fine-tuned baselines.</p> - https://arxiv.org/abs/2511.18659 - Mon, 24 Nov 2025 00:11:14 +0000 + SimWorld: An Open-ended Realistic Simulator for Autonomous Agents in Physical and Social Worlds + https://arxiv.org/abs/2512.01078 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2512.01078.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Jiawei Ren, Yan Zhuang, Xiaokang Ye, Lingjun Mao, Xuhong He, Jianzhi Shen, Mrinaal Dogra, Yiming Liang, Ruixuan Zhang, Tianai Yue, Yiqing Yang, Eric Liu, Ryan Wu, Kevin Benavente, Rajiv Mandya Nagaraju, Muhammad Faayez, Xiyan Zhang, Dhruv Vivek Sharma, Xianrui Zhong, Ziqiao Ma, Tianmin Shu, Zhiting Hu, Lianhui Qin</p><p><b>Upvotes:</b> 32</p><p><b>Summary:</b> While LLM/VLM-powered AI agents have advanced rapidly in math, coding, and computer use, their applications in complex physical and social environments remain challenging. Building agents that can survive and thrive in the real world (for example, by autonomously earning income or running a business) requires massive-scale interaction, reasoning, training, and evaluation across diverse embodied scenarios. However, existing world simulators for such development fall short: they often rely on limited hand-crafted environments, simulate simplified game-like physics and social rules, and lack native support for LLM/VLM agents. We introduce SimWorld, a new simulator built on Unreal Engine 5, designed for developing and evaluating LLM/VLM agents in rich, real-world-like settings. SimWorld offers three core capabilities: (1) realistic, open-ended world simulation, including accurate physical and social dynamics and language-driven procedural environment generation; (2) a rich interface for LLM/VLM agents, with multimodal world inputs and open-vocabulary actions at varying levels of abstraction; and (3) diverse and extensible physical and social reasoning scenarios that are easily customizable by users. We demonstrate SimWorld by deploying frontier LLM agents (e.g., GPT-4o, Gemini-2.5-Flash, Claude-3.5, and DeepSeek-Prover-V2) on long-horizon multi-agent delivery tasks involving strategic cooperation and competition. The results reveal distinct reasoning patterns and limitations across models. We open-source SimWorld and hope it becomes a foundational platform for advancing real-world agent intelligence across disciplines: https://simworld.org.</p> + https://arxiv.org/abs/2512.01078 + Sun, 30 Nov 2025 20:58:13 +0000 - Yo'City: Personalized and Boundless 3D Realistic City Scene Generation via Self-Critic Expansion - https://arxiv.org/abs/2511.18734 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.18734.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Keyang Lu, Sifan Zhou, Hongbin Xu, Gang Xu, Zhifei Yang, Yikai Wang, Zhen Xiao, Jieyi Long, Ming Li</p><p><b>Upvotes:</b> 6</p><p><b>Summary:</b> Realistic 3D city generation is fundamental to a wide range of applications, including virtual reality and digital twins. However, most existing methods rely on training a single diffusion model, which limits their ability to generate personalized and boundless city-scale scenes. In this paper, we present Yo'City, a novel agentic framework that enables user-customized and infinitely expandable 3D city generation by leveraging the reasoning and compositional capabilities of off-the-shelf large models. Specifically, Yo'City first conceptualize the city through a top-down planning strategy that defines a hierarchical "City-District-Grid" structure. The Global Planner determines the overall layout and potential functional districts, while the Local Designer further refines each district with detailed grid-level descriptions. Subsequently, the grid-level 3D generation is achieved through a "produce-refine-evaluate" isometric image synthesis loop, followed by image-to-3D generation. To simulate continuous city evolution, Yo'City further introduces a user-interactive, relationship-guided expansion mechanism, which performs scene graph-based distance- and semantics-aware layout optimization, ensuring spatially coherent city growth. To comprehensively evaluate our method, we construct a diverse benchmark dataset and design six multi-dimensional metrics that assess generation quality from the perspectives of semantics, geometry, texture, and layout. Extensive experiments demonstrate that Yo'City consistently outperforms existing state-of-the-art methods across all evaluation aspects.</p> - https://arxiv.org/abs/2511.18734 - Mon, 24 Nov 2025 04:02:48 +0000 + TRivia: Self-supervised Fine-tuning of Vision-Language Models for Table Recognition + https://arxiv.org/abs/2512.01248 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2512.01248.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Junyuan Zhang, Bin Wang, Qintong Zhang, Fan Wu, Zichen Wen, Jialin Lu, Junjie Shan, Ziqi Zhao, Shuya Yang, Ziling Wang, Ziyang Miao, Huaping Zhong, Yuhang Zang, Xiaoyi Dong, Ka-Ho Chow, Conghui He</p><p><b>Upvotes:</b> 8</p><p><b>Summary:</b> 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</p> + https://arxiv.org/abs/2512.01248 + Mon, 01 Dec 2025 03:49:00 +0000 - HunyuanVideo 1.5 Technical Report - https://arxiv.org/abs/2511.18870 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.18870.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Bing Wu, Chang Zou, Changlin Li, Duojun Huang, Fang Yang, Hao Tan, Jack Peng, Jianbing Wu, Jiangfeng Xiong, Jie Jiang, Linus, Patrol, Peizhen Zhang, Peng Chen, Penghao Zhao, Qi Tian, Songtao Liu, Weijie Kong, Weiyan Wang, Xiao He, Xin Li, Xinchi Deng, Xuefei Zhe, Yang Li, Yanxin Long, Yuanbo Peng, Yue Wu, Yuhong Liu, Zhenyu Wang, Zuozhuo Dai, Bo Peng, Coopers Li, Gu Gong, Guojian Xiao, Jiahe Tian, Jiaxin Lin, Jie Liu, Jihong Zhang, Jiesong Lian, Kaihang Pan, Lei Wang, Lin Niu, Mingtao Chen, Mingyang Chen, Mingzhe Zheng, Miles Yang, Qiangqiang Hu, Qi Yang, Qiuyong Xiao, Runzhou Wu, Ryan Xu, Rui Yuan, Shanshan Sang, Shisheng Huang, Siruis Gong, Shuo Huang, Weiting Guo, Xiang Yuan, Xiaojia Chen, Xiawei Hu, Wenzhi Sun, Xiele Wu, Xianshun Ren, Xiaoyan Yuan, Xiaoyue Mi, Yepeng Zhang, Yifu Sun, Yiting Lu, Yitong Li, You Huang, Yu Tang, Yixuan Li, Yuhang Deng, Yuan Zhou, Zhichao Hu, Zhiguang Liu, Zhihe Yang, Zilin Yang, Zhenzhi Lu, Zixiang Zhou, Zhao Zhong</p><p><b>Upvotes:</b> 21</p><p><b>Summary:</b> We present HunyuanVideo 1.5, a lightweight yet powerful open-source video generation model that achieves state-of-the-art visual quality and motion coherence with only 8.3 billion parameters, enabling efficient inference on consumer-grade GPUs. This achievement is built upon several key components, including meticulous data curation, an advanced DiT architecture featuring selective and sliding tile attention (SSTA), enhanced bilingual understanding through glyph-aware text encoding, progressive pre-training and post-training, and an efficient video super-resolution network. Leveraging these designs, we developed a unified framework capable of high-quality text-to-video and image-to-video generation across multiple durations and resolutions.Extensive experiments demonstrate that this compact and proficient model establishes a new state-of-the-art among open-source video generation models. By releasing the code and model weights, we provide the community with a high-performance foundation that lowers the barrier to video creation and research, making advanced video generation accessible to a broader audience. All open-source assets are publicly available at https://github.com/Tencent-Hunyuan/HunyuanVideo-1.5.</p> - https://arxiv.org/abs/2511.18870 - Mon, 24 Nov 2025 08:22:07 +0000 + InternVideo-Next: Towards General Video Foundation Models without Video-Text Supervision + https://arxiv.org/abs/2512.01342 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2512.01342.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Chenting Wang, Yuhan Zhu, Yicheng Xu, Jiange Yang, Ziang Yan, Yali Wang, Yi Wang, Limin Wang</p><p><b>Upvotes:</b> 14</p><p><b>Summary:</b> 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.</p> + https://arxiv.org/abs/2512.01342 + Mon, 01 Dec 2025 06:57:39 +0000 - MagicWorld: Interactive Geometry-driven Video World Exploration - https://arxiv.org/abs/2511.18886 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.18886.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Guangyuan Li, Siming Zheng, Shuolin Xu, Jinwei Chen, Bo Li, Xiaobin Hu, Lei Zhao, Peng-Tao Jiang</p><p><b>Upvotes:</b> 17</p><p><b>Summary:</b> Recent interactive video world model methods generate scene evolution conditioned on user instructions. Although they achieve impressive results, two key limitations remain. First, they fail to fully exploit the correspondence between instruction-driven scene motion and the underlying 3D geometry, which results in structural instability under viewpoint changes. Second, they easily forget historical information during multi-step interaction, resulting in error accumulation and progressive drift in scene semantics and structure. To address these issues, we propose MagicWorld, an interactive video world model that integrates 3D geometric priors and historical retrieval. MagicWorld starts from a single scene image, employs user actions to drive dynamic scene evolution, and autoregressively synthesizes continuous scenes. We introduce the Action-Guided 3D Geometry Module (AG3D), which constructs a point cloud from the first frame of each interaction and the corresponding action, providing explicit geometric constraints for viewpoint transitions and thereby improving structural consistency. We further propose History Cache Retrieval (HCR) mechanism, which retrieves relevant historical frames during generation and injects them as conditioning signals, helping the model utilize past scene information and mitigate error accumulation. Experimental results demonstrate that MagicWorld achieves notable improvements in scene stability and continuity across interaction iterations.</p> - https://arxiv.org/abs/2511.18886 - Mon, 24 Nov 2025 08:41:28 +0000 + Stabilizing Reinforcement Learning with LLMs: Formulation and Practices + https://arxiv.org/abs/2512.01374 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2512.01374.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Chujie Zheng, Kai Dang, Bowen Yu, Mingze Li, Huiqiang Jiang, Junrong Lin, Yuqiong Liu, An Yang, Jingren Zhou, Junyang Lin</p><p><b>Upvotes:</b> 77</p><p><b>Summary:</b> 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.</p> + https://arxiv.org/abs/2512.01374 + Mon, 01 Dec 2025 07:45:39 +0000 - One4D: Unified 4D Generation and Reconstruction via Decoupled LoRA Control - https://arxiv.org/abs/2511.18922 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.18922.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Zhenxing Mi, Yuxin Wang, Dan Xu</p><p><b>Upvotes:</b> 10</p><p><b>Summary:</b> We present One4D, a unified framework for 4D generation and reconstruction that produces dynamic 4D content as synchronized RGB frames and pointmaps. By consistently handling varying sparsities of conditioning frames through a Unified Masked Conditioning (UMC) mechanism, One4D can seamlessly transition between 4D generation from a single image, 4D reconstruction from a full video, and mixed generation and reconstruction from sparse frames. Our framework adapts a powerful video generation model for joint RGB and pointmap generation, with carefully designed network architectures. The commonly used diffusion finetuning strategies for depthmap or pointmap reconstruction often fail on joint RGB and pointmap generation, quickly degrading the base video model. To address this challenge, we introduce Decoupled LoRA Control (DLC), which employs two modality-specific LoRA adapters to form decoupled computation branches for RGB frames and pointmaps, connected by lightweight, zero-initialized control links that gradually learn mutual pixel-level consistency. Trained on a mixture of synthetic and real 4D datasets under modest computational budgets, One4D produces high-quality RGB frames and accurate pointmaps across both generation and reconstruction tasks. This work represents a step toward general, high-quality geometry-based 4D world modeling using video diffusion models. Project page: https://mizhenxing.github.io/One4D</p> - https://arxiv.org/abs/2511.18922 - Mon, 24 Nov 2025 09:31:23 +0000 + PromptBridge: Cross-Model Prompt Transfer for Large Language Models + https://arxiv.org/abs/2512.01420 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2512.01420.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Yaxuan Wang, Quan Liu, Zhenting Wang, Zichao Li, Wei Wei, Yang Liu, Yujia Bao</p><p><b>Upvotes:</b> 8</p><p><b>Summary:</b> Large language models (LLMs) underpin applications in code generation, mathematical reasoning, and agent-based workflows. In practice, systems access LLMs via commercial APIs or open-source deployments, and the model landscape (e.g., GPT, Claude, Llama) evolves rapidly. This rapid evolution forces frequent model switches driven by capability, cost, deployment constraints, and privacy. Yet prompts are highly model-sensitive: reusing a prompt engineered for one model on another often yields substantially worse performance than a prompt optimized for the target model. We term this phenomenon Model Drifting. Through extensive empirical analysis across diverse LLM configurations, we show that model drifting is both common and severe. To address this challenge, we introduce PromptBridge, a training-free framework that preserves prompt effectiveness under model switches, enabling cross-model prompt transfer without costly per-task or per-model re-optimization. PromptBridge requires only a small set of alignment tasks for calibration. It first applies Model-Adaptive Reflective Prompt Evolution (MAP-RPE) to obtain task- and model-specific optimal prompts via iterative reflective refinement and quantitative evaluation. Using the resulting calibrated prompt pairs for the source and target models, PromptBridge learns a cross-model prompt mapping. At test time, i.e., for an unseen task, given a source-model prompt, this mapping directly produces an optimized prompt for the target model. Experiments in single-agent and multi-agent settings show that PromptBridge consistently improves downstream accuracy while reducing migration effort. The code will be available soon.</p> + https://arxiv.org/abs/2512.01420 + Mon, 01 Dec 2025 08:55:45 +0000 - MIST: Mutual Information Via Supervised Training - https://arxiv.org/abs/2511.18945 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.18945.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> German Gritsai, Megan Richards, Maxime Méloux, Kyunghyun Cho, Maxime Peyrard</p><p><b>Upvotes:</b> 8</p><p><b>Summary:</b> We propose a fully data-driven approach to designing mutual information (MI) estimators. Since any MI estimator is a function of the observed sample from two random variables, we parameterize this function with a neural network (MIST) and train it end-to-end to predict MI values. Training is performed on a large meta-dataset of 625,000 synthetic joint distributions with known ground-truth MI. To handle variable sample sizes and dimensions, we employ a two-dimensional attention scheme ensuring permutation invariance across input samples. To quantify uncertainty, we optimize a quantile regression loss, enabling the estimator to approximate the sampling distribution of MI rather than return a single point estimate. This research program departs from prior work by taking a fully empirical route, trading universal theoretical guarantees for flexibility and efficiency. Empirically, the learned estimators largely outperform classical baselines across sample sizes and dimensions, including on joint distributions unseen during training. The resulting quantile-based intervals are well-calibrated and more reliable than bootstrap-based confidence intervals, while inference is orders of magnitude faster than existing neural baselines. Beyond immediate empirical gains, this framework yields trainable, fully differentiable estimators that can be embedded into larger learning pipelines. Moreover, exploiting MI's invariance to invertible transformations, meta-datasets can be adapted to arbitrary data modalities via normalizing flows, enabling flexible training for diverse target meta-distributions.</p> - https://arxiv.org/abs/2511.18945 - Mon, 24 Nov 2025 09:55:28 +0000 + StreamGaze: Gaze-Guided Temporal Reasoning and Proactive Understanding in Streaming Videos + https://arxiv.org/abs/2512.01707 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2512.01707.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Daeun Lee, Subhojyoti Mukherjee, Branislav Kveton, Ryan A. Rossi, Viet Dac Lai, Seunghyun Yoon, Trung Bui, Franck Dernoncourt, Mohit Bansal</p><p><b>Upvotes:</b> 7</p><p><b>Summary:</b> 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.</p> + https://arxiv.org/abs/2512.01707 + Mon, 01 Dec 2025 14:15:44 +0000 - MedSAM3: Delving into Segment Anything with Medical Concepts - https://arxiv.org/abs/2511.19046 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.19046.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Anglin Liu, Rundong Xue, Xu R. Cao, Yifan Shen, Yi Lu, Xiang Li, Qianqian Chen, Jintai Chen</p><p><b>Upvotes:</b> 47</p><p><b>Summary:</b> Medical image segmentation is fundamental for biomedical discovery. Existing methods lack generalizability and demand extensive, time-consuming manual annotation for new clinical application. Here, we propose MedSAM-3, a text promptable medical segmentation model for medical image and video segmentation. By fine-tuning the Segment Anything Model (SAM) 3 architecture on medical images paired with semantic conceptual labels, our MedSAM-3 enables medical Promptable Concept Segmentation (PCS), allowing precise targeting of anatomical structures via open-vocabulary text descriptions rather than solely geometric prompts. We further introduce the MedSAM-3 Agent, a framework that integrates Multimodal Large Language Models (MLLMs) to perform complex reasoning and iterative refinement in an agent-in-the-loop workflow. Comprehensive experiments across diverse medical imaging modalities, including X-ray, MRI, Ultrasound, CT, and video, demonstrate that our approach significantly outperforms existing specialist and foundation models. We will release our code and model at https://github.com/Joey-S-Liu/MedSAM3.</p> - https://arxiv.org/abs/2511.19046 - Mon, 24 Nov 2025 12:34:38 +0000 + DiG-Flow: Discrepancy-Guided Flow Matching for Robust VLA Models + https://arxiv.org/abs/2512.01715 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2512.01715.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Wanpeng Zhang, Ye Wang, Hao Luo, Haoqi Yuan, Yicheng Feng, Sipeng Zheng, Qin Jin, Zongqing Lu</p><p><b>Upvotes:</b> 9</p><p><b>Summary:</b> 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.</p> + https://arxiv.org/abs/2512.01715 + Mon, 01 Dec 2025 14:21:15 +0000 - DiffSeg30k: A Multi-Turn Diffusion Editing Benchmark for Localized AIGC Detection - https://arxiv.org/abs/2511.19111 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.19111.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Hai Ci, Ziheng Peng, Pei Yang, Yingxin Xuan, Mike Zheng Shou</p><p><b>Upvotes:</b> 3</p><p><b>Summary:</b> Diffusion-based editing enables realistic modification of local image regions, making AI-generated content harder to detect. Existing AIGC detection benchmarks focus on classifying entire images, overlooking the localization of diffusion-based edits. We introduce DiffSeg30k, a publicly available dataset of 30k diffusion-edited images with pixel-level annotations, designed to support fine-grained detection. DiffSeg30k features: 1) In-the-wild images--we collect images or image prompts from COCO to reflect real-world content diversity; 2) Diverse diffusion models--local edits using eight SOTA diffusion models; 3) Multi-turn editing--each image undergoes up to three sequential edits to mimic real-world sequential editing; and 4) Realistic editing scenarios--a vision-language model (VLM)-based pipeline automatically identifies meaningful regions and generates context-aware prompts covering additions, removals, and attribute changes. DiffSeg30k shifts AIGC detection from binary classification to semantic segmentation, enabling simultaneous localization of edits and identification of the editing models. We benchmark three baseline segmentation approaches, revealing significant challenges in semantic segmentation tasks, particularly concerning robustness to image distortions. Experiments also reveal that segmentation models, despite being trained for pixel-level localization, emerge as highly reliable whole-image classifiers of diffusion edits, outperforming established forgery classifiers while showing great potential in cross-generator generalization. We believe DiffSeg30k will advance research in fine-grained localization of AI-generated content by demonstrating the promise and limitations of segmentation-based methods. DiffSeg30k is released at: https://huggingface.co/datasets/Chaos2629/Diffseg30k</p> - https://arxiv.org/abs/2511.19111 - Mon, 24 Nov 2025 13:43:54 +0000 + GR-RL: Going Dexterous and Precise for Long-Horizon Robotic Manipulation + https://arxiv.org/abs/2512.01801 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2512.01801.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Yunfei Li, Xiao Ma, Jiafeng Xu, Yu Cui, Zhongren Cui, Zhigang Han, Liqun Huang, Tao Kong, Yuxiao Liu, Hao Niu, Wanli Peng, Jingchao Qiao, Zeyu Ren, Haixin Shi, Zhi Su, Jiawen Tian, Yuyang Xiao, Shenyu Zhang, Liwei Zheng, Hang Li, Yonghui Wu</p><p><b>Upvotes:</b> 22</p><p><b>Summary:</b> We present GR-RL, a robotic learning framework that turns a generalist vision-language-action (VLA) policy into a highly capable specialist for long-horizon dexterous manipulation. Assuming the optimality of human demonstrations is core to existing VLA policies. However, we claim that in highly dexterous and precise manipulation tasks, human demonstrations are noisy and suboptimal. GR-RL proposes a multi-stage training pipeline that filters, augments, and reinforces the demonstrations by reinforcement learning. First, GR-RL learns a vision-language-conditioned task progress, filters the demonstration trajectories, and only keeps the transitions that contribute positively to the progress. Specifically, we show that by directly applying offline RL with sparse reward, the resulting Q-values can be treated as a robust progress function. Next, we introduce morphological symmetry augmentation that greatly improves the generalization and performance of GR-RL. Lastly, to better align the VLA policy with its deployment behaviors for high-precision control, we perform online RL by learning a latent space noise predictor. With this pipeline, GR-RL is, to our knowledge, the first learning-based policy that can autonomously lace up a shoe by threading shoelaces through multiple eyelets with an 83.3% success rate, a task requiring long-horizon reasoning, millimeter-level precision, and compliant soft-body interaction. We hope GR-RL provides a step toward enabling generalist robot foundations models to specialize into reliable real-world experts.</p> + https://arxiv.org/abs/2512.01801 + Mon, 01 Dec 2025 15:33:59 +0000 - Representational Stability of Truth in Large Language Models - https://arxiv.org/abs/2511.19166 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.19166.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Samantha Dies, Courtney Maynard, Germans Savcisens, Tina Eliassi-Rad</p><p><b>Upvotes:</b> 1</p><p><b>Summary:</b> Large language models (LLMs) are widely used for factual tasks such as "What treats asthma?" or "What is the capital of Latvia?". However, it remains unclear how stably LLMs encode distinctions between true, false, and neither-true-nor-false content in their internal probabilistic representations. We introduce representational stability as the robustness of an LLM's veracity representations to perturbations in the operational definition of truth. We assess representational stability by (i) training a linear probe on an LLM's activations to separate true from not-true statements and (ii) measuring how its learned decision boundary shifts under controlled label changes. Using activations from sixteen open-source models and three factual domains, we compare two types of neither statements. The first are fact-like assertions about entities we believe to be absent from any training data. We call these unfamiliar neither statements. The second are nonfactual claims drawn from well-known fictional contexts. We call these familiar neither statements. The unfamiliar statements induce the largest boundary shifts, producing up to 40% flipped truth judgements in fragile domains (such as word definitions), while familiar fictional statements remain more coherently clustered and yield smaller changes (leq 8.2%). These results suggest that representational stability stems more from epistemic familiarity than from linguistic form. More broadly, our approach provides a diagnostic for auditing and training LLMs to preserve coherent truth assignments under semantic uncertainty, rather than optimizing for output accuracy alone.</p> - https://arxiv.org/abs/2511.19166 - Mon, 24 Nov 2025 14:28:50 +0000 + Envision: Benchmarking Unified Understanding & Generation for Causal World Process Insights + https://arxiv.org/abs/2512.01816 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2512.01816.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Juanxi Tian, Siyuan Li, Conghui He, Lijun Wu, Cheng Tan</p><p><b>Upvotes:</b> 87</p><p><b>Summary:</b> 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.</p> + https://arxiv.org/abs/2512.01816 + Mon, 01 Dec 2025 15:52:31 +0000 - AutoEnv: Automated Environments for Measuring Cross-Environment Agent Learning - https://arxiv.org/abs/2511.19304 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.19304.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Jiayi Zhang, Yiran Peng, Fanqi Kong, Yang Cheng, Yifan Wu, Zhaoyang Yu, Jinyu Xiang, Jianhao Ruan, Jinlin Wang, Maojia Song, HongZhang Liu, Xiangru Tang, Bang Liu, Chenglin Wu, Yuyu Luo</p><p><b>Upvotes:</b> 86</p><p><b>Summary:</b> Humans naturally adapt to diverse environments by learning underlying rules across worlds with different dynamics, observations, and reward structures. In contrast, existing agents typically demonstrate improvements via self-evolving within a single domain, implicitly assuming a fixed environment distribution. Cross-environment learning has remained largely unmeasured: there is no standard collection of controllable, heterogeneous environments, nor a unified way to represent how agents learn. We address these gaps in two steps. First, we propose AutoEnv, an automated framework that treats environments as factorizable distributions over transitions, observations, and rewards, enabling low-cost (4.12 USD on average) generation of heterogeneous worlds. Using AutoEnv, we construct AutoEnv-36, a dataset of 36 environments with 358 validated levels, on which seven language models achieve 12-49% normalized reward, demonstrating the challenge of AutoEnv-36. Second, we formalize agent learning as a component-centric process driven by three stages of Selection, Optimization, and Evaluation applied to an improvable agent component. Using this formulation, we design eight learning methods and evaluate them on AutoEnv-36. Empirically, the gain of any single learning method quickly decrease as the number of environments increases, revealing that fixed learning methods do not scale across heterogeneous environments. Environment-adaptive selection of learning methods substantially improves performance but exhibits diminishing returns as the method space expands. These results highlight both the necessity and the current limitations of agent learning for scalable cross-environment generalization, and position AutoEnv and AutoEnv-36 as a testbed for studying cross-environment agent learning. The code is avaiable at https://github.com/FoundationAgents/AutoEnv.</p> - https://arxiv.org/abs/2511.19304 - Mon, 24 Nov 2025 16:54:23 +0000 + InnoGym: Benchmarking the Innovation Potential of AI Agents + https://arxiv.org/abs/2512.01822 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2512.01822.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Jintian Zhang, Kewei Xu, Jingsheng Zheng, Zhuoyun Yu, Yuqi Zhu, Yujie Luo, Lanning Wei, Shuofei Qiao, Lun Du, Da Zheng, Shumin Deng, Huajun Chen, Ningyu Zhang</p><p><b>Upvotes:</b> 33</p><p><b>Summary:</b> 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.</p> + https://arxiv.org/abs/2512.01822 + Mon, 01 Dec 2025 16:03:04 +0000 - PRInTS: Reward Modeling for Long-Horizon Information Seeking - https://arxiv.org/abs/2511.19314 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.19314.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Jaewoo Lee, Archiki Prasad, Justin Chih-Yao Chen, Zaid Khan, Elias Stengel-Eskin, Mohit Bansal</p><p><b>Upvotes:</b> 5</p><p><b>Summary:</b> Information-seeking is a core capability for AI agents, requiring them to gather and reason over tool-generated information across long trajectories. However, such multi-step information-seeking tasks remain challenging for agents backed by language models. While process reward models (PRMs) can guide agents by ranking candidate steps at test-time, existing PRMs, designed for short reasoning with binary judgment, cannot capture richer dimensions of information-seeking steps, such as tool interactions and reasoning over tool outputs, nor handle the rapidly growing context in long-horizon tasks. To address these limitations, we introduce PRInTS, a generative PRM trained with dual capabilities: (1) dense scoring based on the PRM's reasoning across multiple step quality dimensions (e.g., interpretation of tool outputs, tool call informativeness) and (2) trajectory summarization that compresses the growing context while preserving essential information for step evaluation. Extensive evaluations across FRAMES, GAIA (levels 1-3), and WebWalkerQA (easy-hard) benchmarks on multiple models, along with ablations, reveal that best-of-n sampling with PRInTS enhances information-seeking abilities of open-source models as well as specialized agents, matching or surpassing the performance of frontier models with a much smaller backbone agent and outperforming other strong reward modeling baselines.</p> - https://arxiv.org/abs/2511.19314 - Mon, 24 Nov 2025 17:09:43 +0000 + Rectifying LLM Thought from Lens of Optimization + https://arxiv.org/abs/2512.01925 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2512.01925.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Junnan Liu, Hongwei Liu, Songyang Zhang, Kai Chen</p><p><b>Upvotes:</b> 23</p><p><b>Summary:</b> 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.</p> + https://arxiv.org/abs/2512.01925 + Mon, 01 Dec 2025 17:41:08 +0000 - SteadyDancer: Harmonized and Coherent Human Image Animation with First-Frame Preservation - https://arxiv.org/abs/2511.19320 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.19320.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Jiaming Zhang, Shengming Cao, Rui Li, Xiaotong Zhao, Yutao Cui, Xinglin Hou, Gangshan Wu, Haolan Chen, Yu Xu, Limin Wang, Kai Ma</p><p><b>Upvotes:</b> 38</p><p><b>Summary:</b> Preserving first-frame identity while ensuring precise motion control is a fundamental challenge in human image animation. The Image-to-Motion Binding process of the dominant Reference-to-Video (R2V) paradigm overlooks critical spatio-temporal misalignments common in real-world applications, leading to failures such as identity drift and visual artifacts. We introduce SteadyDancer, an Image-to-Video (I2V) paradigm-based framework that achieves harmonized and coherent animation and is the first to ensure first-frame preservation robustly. Firstly, we propose a Condition-Reconciliation Mechanism to harmonize the two conflicting conditions, enabling precise control without sacrificing fidelity. Secondly, we design Synergistic Pose Modulation Modules to generate an adaptive and coherent pose representation that is highly compatible with the reference image. Finally, we employ a Staged Decoupled-Objective Training Pipeline that hierarchically optimizes the model for motion fidelity, visual quality, and temporal coherence. Experiments demonstrate that SteadyDancer achieves state-of-the-art performance in both appearance fidelity and motion control, while requiring significantly fewer training resources than comparable methods.</p> - https://arxiv.org/abs/2511.19320 - Mon, 24 Nov 2025 17:15:55 +0000 + How Far Are We from Genuinely Useful Deep Research Agents? + https://arxiv.org/abs/2512.01948 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2512.01948.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Dingling Zhang, He Zhu, Jincheng Ren, Kangqi Song, Xinran Zhou, Boyu Feng, Shudong Liu, Jiabin Luo, Weihao Xie, Zhaohui Wang, Tianrui Qin, King Zhu, Yuqing Wang, Qianben Chen, Yuchen Eleanor Jiang, Wei Wang, Jiaheng Liu, Wangchunshu Zhou</p><p><b>Upvotes:</b> 50</p><p><b>Summary:</b> 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.</p> + https://arxiv.org/abs/2512.01948 + Mon, 01 Dec 2025 17:58:59 +0000 - DeCo: Frequency-Decoupled Pixel Diffusion for End-to-End Image Generation - https://arxiv.org/abs/2511.19365 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.19365.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Zehong Ma, Longhui Wei, Shuai Wang, Shiliang Zhang, Qi Tian</p><p><b>Upvotes:</b> 61</p><p><b>Summary:</b> Pixel diffusion aims to generate images directly in pixel space in an end-to-end fashion. This approach avoids the limitations of VAE in the two-stage latent diffusion, offering higher model capacity. Existing pixel diffusion models suffer from slow training and inference, as they usually model both high-frequency signals and low-frequency semantics within a single diffusion transformer (DiT). To pursue a more efficient pixel diffusion paradigm, we propose the frequency-DeCoupled pixel diffusion framework. With the intuition to decouple the generation of high and low frequency components, we leverage a lightweight pixel decoder to generate high-frequency details conditioned on semantic guidance from the DiT. This thus frees the DiT to specialize in modeling low-frequency semantics. In addition, we introduce a frequency-aware flow-matching loss that emphasizes visually salient frequencies while suppressing insignificant ones. Extensive experiments show that DeCo achieves superior performance among pixel diffusion models, attaining FID of 1.62 (256x256) and 2.22 (512x512) on ImageNet, closing the gap with latent diffusion methods. Furthermore, our pretrained text-to-image model achieves a leading overall score of 0.86 on GenEval in system-level comparison. Codes are publicly available at https://github.com/Zehong-Ma/DeCo.</p> - https://arxiv.org/abs/2511.19365 - Mon, 24 Nov 2025 17:59:06 +0000 + Script: Graph-Structured and Query-Conditioned Semantic Token Pruning for Multimodal Large Language Models + https://arxiv.org/abs/2512.01949 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2512.01949.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Zhongyu Yang, Dannong Xu, Wei Pang, Yingfang Yuan</p><p><b>Upvotes:</b> 8</p><p><b>Summary:</b> 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.</p> + https://arxiv.org/abs/2512.01949 + Mon, 01 Dec 2025 17:59:11 +0000 - HunyuanOCR Technical Report - https://arxiv.org/abs/2511.19575 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.19575.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Hunyuan Vision Team, Pengyuan Lyu, Xingyu Wan, Gengluo Li, Shangpin Peng, Weinong Wang, Liang Wu, Huawen Shen, Yu Zhou, Canhui Tang, Qi Yang, Qiming Peng, Bin Luo, Hower Yang, Houwen Peng, Hongming Yang, Senhao Xie, Binghong Wu, Mana Yang, Sergey Wang, Raccoon Liu, Dick Zhu, Jie Jiang, Linus, Han Hu, Chengquan Zhang</p><p><b>Upvotes:</b> 17</p><p><b>Summary:</b> This paper presents HunyuanOCR, a commercial-grade, open-source, and lightweight (1B parameters) Vision-Language Model (VLM) dedicated to OCR tasks. The architecture comprises a Native Vision Transformer (ViT) and a lightweight LLM connected via an MLP adapter. HunyuanOCR demonstrates superior performance, outperforming commercial APIs, traditional pipelines, and larger models (e.g., Qwen3-VL-4B). Specifically, it surpasses current public solutions in perception tasks (Text Spotting, Parsing) and excels in semantic tasks (IE, Text Image Translation), securing first place in the ICDAR 2025 DIMT Challenge (Small Model Track). Furthermore, it achieves state-of-the-art (SOTA) results on OCRBench among VLMs with fewer than 3B parameters. HunyuanOCR achieves breakthroughs in three key aspects: 1) Unifying Versatility and Efficiency: We implement comprehensive support for core capabilities including spotting, parsing, IE, VQA, and translation within a lightweight framework. This addresses the limitations of narrow "OCR expert models" and inefficient "General VLMs". 2) Streamlined End-to-End Architecture: Adopting a pure end-to-end paradigm eliminates dependencies on pre-processing modules (e.g., layout analysis). This fundamentally resolves error propagation common in traditional pipelines and simplifies system deployment. 3) Data-Driven and RL Strategies: We confirm the critical role of high-quality data and, for the first time in the industry, demonstrate that Reinforcement Learning (RL) strategies yield significant performance gains in OCR tasks. HunyuanOCR is officially open-sourced on HuggingFace. We also provide a high-performance deployment solution based on vLLM, placing its production efficiency in the top tier. We hope this model will advance frontier research and provide a solid foundation for industrial applications.</p> - https://arxiv.org/abs/2511.19575 - Mon, 24 Nov 2025 17:59:59 +0000 + TUNA: Taming Unified Visual Representations for Native Unified Multimodal Models + https://arxiv.org/abs/2512.02014 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2512.02014.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Zhiheng Liu, Weiming Ren, Haozhe Liu, Zijian Zhou, Shoufa Chen, Haonan Qiu, Xiaoke Huang, Zhaochong An, Fanny Yang, Aditya Patel, Viktar Atliha, Tony Ng, Xiao Han, Chuyan Zhu, Chenyang Zhang, Ding Liu, Juan-Manuel Perez-Rua, Sen He, Jürgen Schmidhuber, Wenhu Chen, Ping Luo, Wei Liu, Tao Xiang, Jonas Schult, Yuren Cong</p><p><b>Upvotes:</b> 54</p><p><b>Summary:</b> Unified multimodal models (UMMs) aim to jointly perform multimodal understanding and generation within a single framework. We present TUNA, a native UMM that builds a unified continuous visual representation by cascading a VAE encoder with a representation encoder. This unified representation space allows end-to-end processing of images and videos for both understanding and generation tasks. Compared to prior UMMs with decoupled representations, TUNA's unified visual space avoids representation format mismatches introduced by separate encoders, outperforming decoupled alternatives in both understanding and generation. Moreover, we observe that stronger pretrained representation encoders consistently yield better performance across all multimodal tasks, highlighting the importance of the representation encoder. Finally, in this unified setting, jointly training on both understanding and generation data allows the two tasks to benefit from each other rather than interfere. Our extensive experiments on multimodal understanding and generation benchmarks show that TUNA achieves state-of-the-art results in image and video understanding, image and video generation, and image editing, demonstrating the effectiveness and scalability of its unified representation design.</p> + https://arxiv.org/abs/2512.02014 + Mon, 01 Dec 2025 18:59:51 +0000 - DR Tulu: Reinforcement Learning with Evolving Rubrics for Deep Research - https://arxiv.org/abs/2511.19399 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.19399.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Rulin Shao, Akari Asai, Shannon Zejiang Shen, Hamish Ivison, Varsha Kishore, Jingming Zhuo, Xinran Zhao, Molly Park, Samuel G. Finlayson, David Sontag, Tyler Murray, Sewon Min, Pradeep Dasigi, Luca Soldaini, Faeze Brahman, Wen-tau Yih, Tongshuang Wu, Luke Zettlemoyer, Yoon Kim, Hannaneh Hajishirzi, Pang Wei Koh</p><p><b>Upvotes:</b> 48</p><p><b>Summary:</b> Deep research models perform multi-step research to produce long-form, well-attributed answers. However, most open deep research models are trained on easily verifiable short-form QA tasks via reinforcement learning with verifiable rewards (RLVR), which does not extend to realistic long-form tasks. We address this with Reinforcement Learning with Evolving Rubrics (RLER), in which we construct and maintain rubrics that co-evolve with the policy model during training; this allows the rubrics to incorporate information that the model has newly explored and to provide discriminative, on-policy feedback. Using RLER, we develop Deep Research Tulu (DR Tulu-8B), the first open model that is directly trained for open-ended, long-form deep research. Across four long-form deep research benchmarks in science, healthcare and general domains, DR Tulu substantially outperforms existing open deep research models, and matches or exceeds proprietary deep research systems, while being significantly smaller and cheaper per query. To facilitate future research, we release all data, models, and code, including our new MCP-based agent infrastructure for deep research systems.</p> - https://arxiv.org/abs/2511.19399 - Mon, 24 Nov 2025 18:35:54 +0000 + Skywork-R1V4: Toward Agentic Multimodal Intelligence through Interleaved Thinking with Images and DeepResearch + https://arxiv.org/abs/2512.02395 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2512.02395.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Yifan Zhang, Liang Hu, Haofeng Sun, Peiyu Wang, Yichen Wei, Shukang Yin, Jiangbo Pei, Wei Shen, Peng Xia, Yi Peng, Tianyidan Xie, Eric Li, Yang Liu, Xuchen Song, Yahui Zhou</p><p><b>Upvotes:</b> 43</p><p><b>Summary:</b> 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.</p> + https://arxiv.org/abs/2512.02395 + Tue, 02 Dec 2025 04:12:57 +0000 - In-Video Instructions: Visual Signals as Generative Control - https://arxiv.org/abs/2511.19401 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.19401.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Gongfan Fang, Xinyin Ma, Xinchao Wang</p><p><b>Upvotes:</b> 28</p><p><b>Summary:</b> Large-scale video generative models have recently demonstrated strong visual capabilities, enabling the prediction of future frames that adhere to the logical and physical cues in the current observation. In this work, we investigate whether such capabilities can be harnessed for controllable image-to-video generation by interpreting visual signals embedded within the frames as instructions, a paradigm we term In-Video Instruction. In contrast to prompt-based control, which provides textual descriptions that are inherently global and coarse, In-Video Instruction encodes user guidance directly into the visual domain through elements such as overlaid text, arrows, or trajectories. This enables explicit, spatial-aware, and unambiguous correspondences between visual subjects and their intended actions by assigning distinct instructions to different objects. Extensive experiments on three state-of-the-art generators, including Veo 3.1, Kling 2.5, and Wan 2.2, show that video models can reliably interpret and execute such visually embedded instructions, particularly in complex multi-object scenarios.</p> - https://arxiv.org/abs/2511.19401 - Mon, 24 Nov 2025 18:38:45 +0000 + WorldMM: Dynamic Multimodal Memory Agent for Long Video Reasoning + https://arxiv.org/abs/2512.02425 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2512.02425.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Woongyeong Yeo, Kangsan Kim, Jaehong Yoon, Sung Ju Hwang</p><p><b>Upvotes:</b> 22</p><p><b>Summary:</b> Recent advances in video large language models have demonstrated strong capabilities in understanding short clips. However, scaling them to hours- or days-long videos remains highly challenging due to limited context capacity and the loss of critical visual details during abstraction. Existing memory-augmented methods mitigate this by leveraging textual summaries of video segments, yet they heavily rely on text and fail to utilize visual evidence when reasoning over complex scenes. Moreover, retrieving from fixed temporal scales further limits their flexibility in capturing events that span variable durations. To address this, we introduce WorldMM, a novel multimodal memory agent that constructs and retrieves from multiple complementary memories, encompassing both textual and visual representations. WorldMM comprises three types of memory: episodic memory indexes factual events across multiple temporal scales, semantic memory continuously updates high-level conceptual knowledge, and visual memory preserves detailed information about scenes. During inference, an adaptive retrieval agent iteratively selects the most relevant memory source and leverages multiple temporal granularities based on the query, continuing until it determines that sufficient information has been gathered. WorldMM significantly outperforms existing baselines across five long video question-answering benchmarks, achieving an average 8.4% performance gain over previous state-of-the-art methods, showing its effectiveness on long video reasoning.</p> + https://arxiv.org/abs/2512.02425 + Tue, 02 Dec 2025 05:14:52 +0000 - UniGame: Turning a Unified Multimodal Model Into Its Own Adversary - https://arxiv.org/abs/2511.19413 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.19413.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Zhaolong Su, Wang Lu, Hao Chen, Sharon Li, Jindong Wang</p><p><b>Upvotes:</b> 17</p><p><b>Summary:</b> Unified Multimodal Models (UMMs) have shown impressive performance in both understanding and generation with a single architecture. However, UMMs still exhibit a fundamental inconsistency: understanding favors compact embeddings, whereas generation favors reconstruction-rich representations. This structural trade-off produces misaligned decision boundaries, degraded cross-modal coherence, and heightened vulnerability under distributional and adversarial shifts. In this paper, we present UniGame, a self-adversarial post-training framework that directly targets the inconsistencies. By applying a lightweight perturber at the shared token interface, UniGame enables the generation branch to actively seek and challenge fragile understanding, turning the model itself into its own adversary. Experiments demonstrate that UniGame significantly improves the consistency (+4.6%). Moreover, it also achieves substantial improvements in understanding (+3.6%), generation (+0.02), out-of-distribution and adversarial robustness (+4.8% and +6.2% on NaturalBench and AdVQA). The framework is architecture-agnostic, introduces less than 1% additional parameters, and is complementary to existing post-training methods. These results position adversarial self-play as a general and effective principle for enhancing the coherence, stability, and unified competence of future multimodal foundation models. The official code is available at: https://github.com/AIFrontierLab/UniGame</p> - https://arxiv.org/abs/2511.19413 - Mon, 24 Nov 2025 18:50:01 +0000 + Does Hearing Help Seeing? Investigating Audio-Video Joint Denoising for Video Generation + https://arxiv.org/abs/2512.02457 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2512.02457.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Jianzong Wu, Hao Lian, Dachao Hao, Ye Tian, Qingyu Shi, Biaolong Chen, Hao Jiang</p><p><b>Upvotes:</b> 12</p><p><b>Summary:</b> 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.</p> + https://arxiv.org/abs/2512.02457 + Tue, 02 Dec 2025 06:31:38 +0000 - Chain-of-Visual-Thought: Teaching VLMs to See and Think Better with Continuous Visual Tokens - https://arxiv.org/abs/2511.19418 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.19418.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Yiming Qin, Bomin Wei, Jiaxin Ge, Konstantinos Kallidromitis, Stephanie Fu, Trevor Darrell, Xudong Wang</p><p><b>Upvotes:</b> 24</p><p><b>Summary:</b> Vision-Language Models (VLMs) excel at reasoning in linguistic space but struggle with perceptual understanding that requires dense visual perception, e.g., spatial reasoning and geometric awareness. This limitation stems from the fact that current VLMs have limited mechanisms to capture dense visual information across spatial dimensions. We introduce Chain-of-Visual-Thought (COVT), a framework that enables VLMs to reason not only in words but also through continuous visual tokens-compact latent representations that encode rich perceptual cues. Within a small budget of roughly 20 tokens, COVT distills knowledge from lightweight vision experts, capturing complementary properties such as 2D appearance, 3D geometry, spatial layout, and edge structure. During training, the VLM with COVT autoregressively predicts these visual tokens to reconstruct dense supervision signals (e.g., depth, segmentation, edges, and DINO features). At inference, the model reasons directly in the continuous visual token space, preserving efficiency while optionally decoding dense predictions for interpretability. Evaluated across more than ten diverse perception benchmarks, including CV-Bench, MMVP, RealWorldQA, MMStar, WorldMedQA, and HRBench, integrating COVT into strong VLMs such as Qwen2.5-VL and LLaVA consistently improves performance by 3% to 16% and demonstrates that compact continuous visual thinking enables more precise, grounded, and interpretable multimodal intelligence.</p> - https://arxiv.org/abs/2511.19418 - Mon, 24 Nov 2025 18:55:19 +0000 + Guided Self-Evolving LLMs with Minimal Human Supervision + https://arxiv.org/abs/2512.02472 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2512.02472.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Wenhao Yu, Zhenwen Liang, Chengsong Huang, Kishan Panaganti, Tianqing Fang, Haitao Mi, Dong Yu</p><p><b>Upvotes:</b> 47</p><p><b>Summary:</b> 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.</p> + https://arxiv.org/abs/2512.02472 + Tue, 02 Dec 2025 07:06:11 +0000 - Flow Map Distillation Without Data - https://arxiv.org/abs/2511.19428 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.19428.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Shangyuan Tong, Nanye Ma, Saining Xie, Tommi Jaakkola</p><p><b>Upvotes:</b> 3</p><p><b>Summary:</b> State-of-the-art flow models achieve remarkable quality but require slow, iterative sampling. To accelerate this, flow maps can be distilled from pre-trained teachers, a procedure that conventionally requires sampling from an external dataset. We argue that this data-dependency introduces a fundamental risk of Teacher-Data Mismatch, as a static dataset may provide an incomplete or even misaligned representation of the teacher's full generative capabilities. This leads us to question whether this reliance on data is truly necessary for successful flow map distillation. In this work, we explore a data-free alternative that samples only from the prior distribution, a distribution the teacher is guaranteed to follow by construction, thereby circumventing the mismatch risk entirely. To demonstrate the practical viability of this philosophy, we introduce a principled framework that learns to predict the teacher's sampling path while actively correcting for its own compounding errors to ensure high fidelity. Our approach surpasses all data-based counterparts and establishes a new state-of-the-art by a significant margin. Specifically, distilling from SiT-XL/2+REPA, our method reaches an impressive FID of 1.45 on ImageNet 256x256, and 1.49 on ImageNet 512x512, both with only 1 sampling step. We hope our work establishes a more robust paradigm for accelerating generative models and motivates the broader adoption of flow map distillation without data.</p> - https://arxiv.org/abs/2511.19428 - Mon, 24 Nov 2025 18:58:55 +0000 + CUDA-L2: Surpassing cuBLAS Performance for Matrix Multiplication through Reinforcement Learning + https://arxiv.org/abs/2512.02551 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2512.02551.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Songqiao Su, Xiaofei Sun, Xiaoya Li, Albert Wang, Jiwei Li, Chris Shum</p><p><b>Upvotes:</b> 11</p><p><b>Summary:</b> 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</p> + https://arxiv.org/abs/2512.02551 + Tue, 02 Dec 2025 09:20:15 +0000 - Cook and Clean Together: Teaching Embodied Agents for Parallel Task Execution - https://arxiv.org/abs/2511.19430 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.19430.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Dingkang Liang, Cheng Zhang, Xiaopeng Xu, Jianzhong Ju, Zhenbo Luo, Xiang Bai</p><p><b>Upvotes:</b> 7</p><p><b>Summary:</b> Task scheduling is critical for embodied AI, enabling agents to follow natural language instructions and execute actions efficiently in 3D physical worlds. However, existing datasets often simplify task planning by ignoring operations research (OR) knowledge and 3D spatial grounding. In this work, we propose Operations Research knowledge-based 3D Grounded Task Scheduling (ORS3D), a new task that requires the synergy of language understanding, 3D grounding, and efficiency optimization. Unlike prior settings, ORS3D demands that agents minimize total completion time by leveraging parallelizable subtasks, e.g., cleaning the sink while the microwave operates. To facilitate research on ORS3D, we construct ORS3D-60K, a large-scale dataset comprising 60K composite tasks across 4K real-world scenes. Furthermore, we propose GRANT, an embodied multi-modal large language model equipped with a simple yet effective scheduling token mechanism to generate efficient task schedules and grounded actions. Extensive experiments on ORS3D-60K validate the effectiveness of GRANT across language understanding, 3D grounding, and scheduling efficiency. The code is available at https://github.com/H-EmbodVis/GRANT</p> - https://arxiv.org/abs/2511.19430 - Mon, 24 Nov 2025 18:59:17 +0000 + DeepSeek-V3.2: Pushing the Frontier of Open Large Language Models + https://arxiv.org/abs/2512.02556 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2512.02556.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> 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</p><p><b>Upvotes:</b> 172</p><p><b>Summary:</b> 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.</p> + https://arxiv.org/abs/2512.02556 + Tue, 02 Dec 2025 09:25:14 +0000 - Fara-7B: An Efficient Agentic Model for Computer Use - https://arxiv.org/abs/2511.19663 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.19663.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Ahmed Awadallah, Yash Lara, Raghav Magazine, Hussein Mozannar, Akshay Nambi, Yash Pandya, Aravind Rajeswaran, Corby Rosset, Alexey Taymanov, Vibhav Vineet, Spencer Whitehead, Andrew Zhao</p><p><b>Upvotes:</b> 8</p><p><b>Summary:</b> Progress in computer use agents (CUAs) has been constrained by the absence of large and high-quality datasets that capture how humans interact with a computer. While LLMs have thrived on abundant textual data, no comparable corpus exists for CUA trajectories. To address these gaps, we introduce FaraGen, a novel synthetic data generation system for multi-step web tasks. FaraGen can propose diverse tasks from frequently used websites, generate multiple solution attempts, and filter successful trajectories using multiple verifiers. It achieves high throughput, yield, and diversity for multi-step web tasks, producing verified trajectories at approximately $1 each. We use this data to train Fara-7B, a native CUA model that perceives the computer using only screenshots, executes actions via predicted coordinates, and is small enough to run on-device. We find that Fara-7B outperforms other CUA models of comparable size on benchmarks like WebVoyager, Online-Mind2Web, and WebTailBench -- our novel benchmark that better captures under-represented web tasks in pre-existing benchmarks. Furthermore, Fara-7B is competitive with much larger frontier models, illustrating key benefits of scalable data generation systems in advancing small efficient agentic models. We are making Fara-7B open-weight on Microsoft Foundry and HuggingFace, and we are releasing WebTailBench.</p> - https://arxiv.org/abs/2511.19663 - Mon, 24 Nov 2025 19:56:28 +0000 + PaperDebugger: A Plugin-Based Multi-Agent System for In-Editor Academic Writing, Review, and Editing + https://arxiv.org/abs/2512.02589 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2512.02589.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Junyi Hou, Andre Lin Huikai, Nuo Chen, Yiwei Gong, Bingsheng He</p><p><b>Upvotes:</b> 38</p><p><b>Summary:</b> 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.</p> + https://arxiv.org/abs/2512.02589 + Tue, 02 Dec 2025 10:00:37 +0000 - What does it mean to understand language? - https://arxiv.org/abs/2511.19757 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.19757.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Colton Casto, Anna Ivanova, Evelina Fedorenko, Nancy Kanwisher</p><p><b>Upvotes:</b> 8</p><p><b>Summary:</b> Language understanding entails not just extracting the surface-level meaning of the linguistic input, but constructing rich mental models of the situation it describes. Here we propose that because processing within the brain's core language system is fundamentally limited, deeply understanding language requires exporting information from the language system to other brain regions that compute perceptual and motor representations, construct mental models, and store our world knowledge and autobiographical memories. We review the existing evidence for this hypothesis, and argue that recent progress in cognitive neuroscience provides both the conceptual foundation and the methods to directly test it, thus opening up a new strategy to reveal what it means, cognitively and neurally, to understand language.</p> - https://arxiv.org/abs/2511.19757 - Mon, 24 Nov 2025 22:21:55 +0000 + RULER-Bench: Probing Rule-based Reasoning Abilities of Next-level Video Generation Models for Vision Foundation Intelligence + https://arxiv.org/abs/2512.02622 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2512.02622.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Xuming He, Zehao Fan, Hengjia Li, Fan Zhuo, Hankun Xu, Senlin Cheng, Di Weng, Haifeng Liu, Can Ye, Boxi Wu</p><p><b>Upvotes:</b> 8</p><p><b>Summary:</b> Recent advances in video generation have enabled the synthesis of videos with strong temporal consistency and impressive visual quality, marking a crucial step toward vision foundation models. To evaluate these video generation models, existing benchmarks primarily focus on factors related to visual perception and understanding, like visual aesthetics, instruction adherence, and temporal coherence. However, the rule-based reasoning capabilities of video generation models remain largely unexplored. Although recent studies have carried out preliminary explorations into whether video models can serve as zero-shot learners, they still lack a fine-grained decomposition of reasoning capabilities and a comprehensive evaluation protocol. To address this gap, we introduce RULER-Bench, a benchmark designed to evaluate the reasoning ability of video generation models from the perspective of cognitive rules. Built upon two fundamental paradigms: text-to-video and image-to-video, RULER-Bench covers 40 representative tasks spanning six rule categories with 622 high-quality annotated instances. For the evaluation of each generated video, we construct a checklist covering four metrics and leverage GPT-o3 to assign scores to each question, achieving 85% alignment with human judgements. Extensive experiments show that the state-of-the-art model achieves only 48.87% on the rule coherence metric, highlighting significant room for improvement in the reasoning capability of next-level video models. We expect that the insight obtained from RULER-Bench will facilitate further development of reasoning-aware video generation, advancing video generation models toward vision foundation intelligence.</p> + https://arxiv.org/abs/2512.02622 + Tue, 02 Dec 2025 10:29:51 +0000 - Scaling Agentic Reinforcement Learning for Tool-Integrated Reasoning in VLMs - https://arxiv.org/abs/2511.19773 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.19773.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Meng Lu, Ran Xu, Yi Fang, Wenxuan Zhang, Yue Yu, Gaurav Srivastava, Yuchen Zhuang, Mohamed Elhoseiny, Charles Fleming, Carl Yang, Zhengzhong Tu, Yang Xie, Guanghua Xiao, Hanrui Wang, Di Jin, Wenqi Shi, Xuan Wang</p><p><b>Upvotes:</b> 9</p><p><b>Summary:</b> While recent vision-language models (VLMs) demonstrate strong image understanding, their ability to "think with images", i.e., to reason through multi-step visual interactions, remains limited. We introduce VISTA-Gym, a scalable training environment for incentivizing tool-integrated visual reasoning capabilities in VLMs. VISTA-Gym unifies diverse real-world multimodal reasoning tasks (7 tasks from 13 datasets in total) with a standardized interface for visual tools (e.g., grounding, parsing), executable interaction loops, verifiable feedback signals, and efficient trajectory logging, enabling visual agentic reinforcement learning at scale. While recent VLMs exhibit strong text-only reasoning, both proprietary and open-source models still struggle with tool selection, invocation, and coordination. With VISTA-Gym, we train VISTA-R1 to interleave tool-use with agentic reasoning via multi-turn trajectory sampling and end-to-end reinforcement learning. Extensive experiments across 11 public reasoning-intensive VQA benchmarks show that VISTA-R1-8B outperforms state-of-the-art baselines with similar sizes by 9.51%-18.72%, demonstrating VISTA-Gym as an effective training ground to unlock the tool-integrated reasoning capabilities for VLMs.</p> - https://arxiv.org/abs/2511.19773 - Mon, 24 Nov 2025 22:58:26 +0000 + SeeNav-Agent: Enhancing Vision-Language Navigation with Visual Prompt and Step-Level Policy Optimization + https://arxiv.org/abs/2512.02631 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2512.02631.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Zhengcheng Wang, Zichuan Lin, Yijun Yang, Haobo Fu, Deheng Ye</p><p><b>Upvotes:</b> 6</p><p><b>Summary:</b> Existing Vision-Language Navigation (VLN) agents based on Large Vision-Language Models (LVLMs) often suffer from perception errors, reasoning errors, and planning errors, which significantly hinder their navigation performance. To address these limitations, a novel VLN agent framework, named SeeNav-Agent, is proposed in this work. First, to reduce perception hallucinations of the visual module of the VLN agent, a dual-view Visual Prompt (VP) technique is introduced in the input space, which can also improve the agent's understanding of current spatial states. Subsequently, a novel step-level Reinforcement Fine-Tuning (RFT) method, Step Reward Group Policy Optimization (SRGPO), is designed for the post-training of VLN agents. In SRGPO, we first define verifiable process rewards for the navigation task, and then perform efficient step-level advantage estimation by randomly grouping different navigation steps. SRGPO provides dense reward signals for the reinforcement learning process of the VLN agent and enhances its planning capability. Experimental results on the EmbodiedBench Navigation benchmark indicate that by introducing the zero-shot VP module, the GPT-4.1 achieves a navigation success rate of 86.7%, surpassing the current best LVLM by approximately 20 percentage points (pp). Through post-training based on SRGPO, the Qwen2.5-VL-3B model reaches a navigation success rate of 72.3%, outperforming the best existing LVLM model by 5.6 pp. Moreover, compared to RFT algorithms such as GRPO and GiGPO, the proposed SRGPO demonstrates significant improvements in training stability, convergence efficiency, and generalization capability.</p> + https://arxiv.org/abs/2512.02631 + Tue, 02 Dec 2025 10:40:46 +0000 - Terminal Velocity Matching - https://arxiv.org/abs/2511.19797 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.19797.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Linqi Zhou, Mathias Parger, Ayaan Haque, Jiaming Song</p><p><b>Upvotes:</b> 11</p><p><b>Summary:</b> We propose Terminal Velocity Matching (TVM), a generalization of flow matching that enables high-fidelity one- and few-step generative modeling. TVM models the transition between any two diffusion timesteps and regularizes its behavior at its terminal time rather than at the initial time. We prove that TVM provides an upper bound on the 2-Wasserstein distance between data and model distributions when the model is Lipschitz continuous. However, since Diffusion Transformers lack this property, we introduce minimal architectural changes that achieve stable, single-stage training. To make TVM efficient in practice, we develop a fused attention kernel that supports backward passes on Jacobian-Vector Products, which scale well with transformer architectures. On ImageNet-256x256, TVM achieves 3.29 FID with a single function evaluation (NFE) and 1.99 FID with 4 NFEs. It similarly achieves 4.32 1-NFE FID and 2.94 4-NFE FID on ImageNet-512x512, representing state-of-the-art performance for one/few-step models from scratch.</p> - https://arxiv.org/abs/2511.19797 - Mon, 24 Nov 2025 23:55:45 +0000 + SR-GRPO: Stable Rank as an Intrinsic Geometric Reward for Large Language Model Alignment + https://arxiv.org/abs/2512.02807 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2512.02807.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Yixuan Tang, Yi Yang</p><p><b>Upvotes:</b> 7</p><p><b>Summary:</b> 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.</p> + https://arxiv.org/abs/2512.02807 + Tue, 02 Dec 2025 14:21:29 +0000 - ReDirector: Creating Any-Length Video Retakes with Rotary Camera Encoding - https://arxiv.org/abs/2511.19827 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.19827.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Byeongjun Park, Byung-Hoon Kim, Hyungjin Chung, Jong Chul Ye</p><p><b>Upvotes:</b> 11</p><p><b>Summary:</b> We present ReDirector, a novel camera-controlled video retake generation method for dynamically captured variable-length videos. In particular, we rectify a common misuse of RoPE in previous works by aligning the spatiotemporal positions of the input video and the target retake. Moreover, we introduce Rotary Camera Encoding (RoCE), a camera-conditioned RoPE phase shift that captures and integrates multi-view relationships within and across the input and target videos. By integrating camera conditions into RoPE, our method generalizes to out-of-distribution camera trajectories and video lengths, yielding improved dynamic object localization and static background preservation. Extensive experiments further demonstrate significant improvements in camera controllability, geometric consistency, and video quality across various trajectories and lengths.</p> - https://arxiv.org/abs/2511.19827 - Tue, 25 Nov 2025 01:38:56 +0000 + Steering Vision-Language-Action Models as Anti-Exploration: A Test-Time Scaling Approach + https://arxiv.org/abs/2512.02834 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2512.02834.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Siyuan Yang, Yang Zhang, Haoran He, Ling Pan, Xiu Li, Chenjia Bai, Xuelong Li</p><p><b>Upvotes:</b> 37</p><p><b>Summary:</b> 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 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.</p> + https://arxiv.org/abs/2512.02834 + Tue, 02 Dec 2025 14:42:54 +0000 - Inferix: A Block-Diffusion based Next-Generation Inference Engine for World Simulation - https://arxiv.org/abs/2511.20714 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.20714.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Inferix Team, Tianyu Feng, Yizeng Han, Jiahao He, Yuanyu He, Xi Lin, Teng Liu, Hanfeng Lu, Jiasheng Tang, Wei Wang, Zhiyuan Wang, Jichao Wu, Mingyang Yang, Yinghao Yu, Zeyu Zhang, Bohan Zhuang</p><p><b>Upvotes:</b> 41</p><p><b>Summary:</b> World models serve as core simulators for fields such as agentic AI, embodied AI, and gaming, capable of generating long, physically realistic, and interactive high-quality videos. Moreover, scaling these models could unlock emergent capabilities in visual perception, understanding, and reasoning, paving the way for a new paradigm that moves beyond current LLM-centric vision foundation models. A key breakthrough empowering them is the semi-autoregressive (block-diffusion) decoding paradigm, which merges the strengths of diffusion and autoregressive methods by generating video tokens in block-applying diffusion within each block while conditioning on previous ones, resulting in more coherent and stable video sequences. Crucially, it overcomes limitations of standard video diffusion by reintroducing LLM-style KV Cache management, enabling efficient, variable-length, and high-quality generation. Therefore, Inferix is specifically designed as a next-generation inference engine to enable immersive world synthesis through optimized semi-autoregressive decoding processes. This dedicated focus on world simulation distinctly sets it apart from systems engineered for high-concurrency scenarios (like vLLM or SGLang) and from classic video diffusion models (such as xDiTs). Inferix further enhances its offering with interactive video streaming and profiling, enabling real-time interaction and realistic simulation to accurately model world dynamics. Additionally, it supports efficient benchmarking through seamless integration of LV-Bench, a new fine-grained evaluation benchmark tailored for minute-long video generation scenarios. We hope the community will work together to advance Inferix and foster world model exploration.</p> - https://arxiv.org/abs/2511.20714 - Tue, 25 Nov 2025 01:45:04 +0000 + Glance: Accelerating Diffusion Models with 1 Sample + https://arxiv.org/abs/2512.02899 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2512.02899.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Zhuobai Dong, Rui Zhao, Songjie Wu, Junchao Yi, Linjie Li, Zhengyuan Yang, Lijuan Wang, Alex Jinpeng Wang</p><p><b>Upvotes:</b> 23</p><p><b>Summary:</b> 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.</p> + https://arxiv.org/abs/2512.02899 + Tue, 02 Dec 2025 16:05:21 +0000 - GigaWorld-0: World Models as Data Engine to Empower Embodied AI - https://arxiv.org/abs/2511.19861 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.19861.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> GigaWorld Team, Angen Ye, Boyuan Wang, Chaojun Ni, Guan Huang, Guosheng Zhao, Haoyun Li, Jiagang Zhu, Kerui Li, Mengyuan Xu, Qiuping Deng, Siting Wang, Wenkang Qin, Xinze Chen, Xiaofeng Wang, Yankai Wang, Yu Cao, Yifan Chang, Yuan Xu, Yun Ye, Yang Wang, Yukun Zhou, Zhengyuan Zhang, Zhehao Dong, Zheng Zhu</p><p><b>Upvotes:</b> 27</p><p><b>Summary:</b> World models are emerging as a foundational paradigm for scalable, data-efficient embodied AI. In this work, we present GigaWorld-0, a unified world model framework designed explicitly as a data engine for Vision-Language-Action (VLA) learning. GigaWorld-0 integrates two synergistic components: GigaWorld-0-Video, which leverages large-scale video generation to produce diverse, texture-rich, and temporally coherent embodied sequences under fine-grained control of appearance, camera viewpoint, and action semantics; and GigaWorld-0-3D, which combines 3D generative modeling, 3D Gaussian Splatting reconstruction, physically differentiable system identification, and executable motion planning to ensure geometric consistency and physical realism. Their joint optimization enables the scalable synthesis of embodied interaction data that is visually compelling, spatially coherent, physically plausible, and instruction-aligned. Training at scale is made feasible through our efficient GigaTrain framework, which exploits FP8-precision and sparse attention to drastically reduce memory and compute requirements. We conduct comprehensive evaluations showing that GigaWorld-0 generates high-quality, diverse, and controllable data across multiple dimensions. Critically, VLA model (e.g., GigaBrain-0) trained on GigaWorld-0-generated data achieve strong real-world performance, significantly improving generalization and task success on physical robots without any real-world interaction during training.</p> - https://arxiv.org/abs/2511.19861 - Tue, 25 Nov 2025 03:00:42 +0000 + DynamicVerse: A Physically-Aware Multimodal Framework for 4D World Modeling + https://arxiv.org/abs/2512.03000 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2512.03000.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Kairun Wen, Yuzhi Huang, Runyu Chen, Hui Zheng, Yunlong Lin, Panwang Pan, Chenxin Li, Wenyan Cong, Jian Zhang, Junbin Lu, Chenguo Lin, Dilin Wang, Zhicheng Yan, Hongyu Xu, Justin Theiss, Yue Huang, Xinghao Ding, Rakesh Ranjan, Zhiwen Fan</p><p><b>Upvotes:</b> 25</p><p><b>Summary:</b> 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.</p> + https://arxiv.org/abs/2512.03000 + Tue, 02 Dec 2025 18:24:27 +0000 - Agent0-VL: Exploring Self-Evolving Agent for Tool-Integrated Vision-Language Reasoning - https://arxiv.org/abs/2511.19900 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.19900.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Jiaqi Liu, Kaiwen Xiong, Peng Xia, Yiyang Zhou, Haonian Ji, Lu Feng, Siwei Han, Mingyu Ding, Huaxiu Yao</p><p><b>Upvotes:</b> 46</p><p><b>Summary:</b> Vision-language agents have achieved remarkable progress in a variety of multimodal reasoning tasks; however, their learning remains constrained by the limitations of human-annotated supervision. Recent self-rewarding approaches attempt to overcome this constraint by allowing models to act as their own critics or reward providers. Yet, purely text-based self-evaluation struggles to verify complex visual reasoning steps and often suffers from evaluation hallucinations. To address these challenges, inspired by recent advances in tool-integrated reasoning, we propose Agent0-VL, a self-evolving vision-language agent that achieves continual improvement with tool-integrated reasoning. Agent0-VL incorporates tool usage not only into reasoning but also into self-evaluation and self-repair, enabling the model to introspect, verify, and refine its reasoning through evidence-grounded analysis. It unifies two synergistic roles within a single LVLM: a Solver that performs multi-turn tool-integrated reasoning, and a Verifier that generates structured feedback and fine-grained self-rewards through tool-grounded critique. These roles interact through a Self-Evolving Reasoning Cycle, where tool-based verification and reinforcement learning jointly align the reasoning and evaluation distributions for stable self-improvement. Through this zero-external-reward evolution, Agent0-VL aligns its reasoning and verification behaviors without any human annotation or external reward models, achieving continual self-improvement. Experiments on geometric problem solving and visual scientific analysis show that Agent0-VL achieves an 12.5% improvement over the base model. Our code is available at https://github.com/aiming-lab/Agent0/Agent0-VL{this https URL}.</p> - https://arxiv.org/abs/2511.19900 - Tue, 25 Nov 2025 04:15:14 +0000 + ViSAudio: End-to-End Video-Driven Binaural Spatial Audio Generation + https://arxiv.org/abs/2512.03036 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2512.03036.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Mengchen Zhang, Qi Chen, Tong Wu, Zihan Liu, Dahua Lin</p><p><b>Upvotes:</b> 20</p><p><b>Summary:</b> 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.</p> + https://arxiv.org/abs/2512.03036 + Tue, 02 Dec 2025 18:56:12 +0000 - SSA: Sparse Sparse Attention by Aligning Full and Sparse Attention Outputs in Feature Space - https://arxiv.org/abs/2511.20102 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.20102.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Zhenyi Shen, Junru Lu, Lin Gui, Jiazheng Li, Yulan He, Di Yin, Xing Sun</p><p><b>Upvotes:</b> 24</p><p><b>Summary:</b> The quadratic complexity of full attention limits efficient long-context processing in large language models (LLMs). Sparse attention mitigates this cost by restricting each query to attend to a subset of previous tokens; however, training-free approaches often lead to severe performance degradation. Native sparse-attention methods (e.g., NSA, MoBA) alleviate this issue, yet exhibit a critical paradox: they produce lower attention sparsity than full-attention models, despite aiming to approximate full attention, which may constrain their effectiveness. We attribute this paradox to gradient update deficiency: low-ranked key-value pairs excluded during sparse training receive neither forward contribution nor backward gradients, and thus never learn proper suppression. To overcome this limitation, we propose SSA (Sparse Sparse Attention), a unified training framework that considers both sparse and full attention and enforces bidirectional alignment at every layer. This design preserves gradient flow to all tokens while explicitly encouraging sparse-attention outputs to align with their full-attention counterparts, thereby promoting stronger sparsity. As a result, SSA achieves state-of-the-art performance under both sparse and full attention inference across multiple commonsense benchmarks. Furthermore, SSA enables models to adapt smoothly to varying sparsity budgets; performance improves consistently as more tokens are allowed to attend, supporting flexible compute-performance trade-offs at inference time. Finally, we show that native sparse-attention training surprisingly improves long-context extrapolation by mitigating the over-allocation of attention values in sink areas, with SSA demonstrating the strongest extrapolation capability.</p> - https://arxiv.org/abs/2511.20102 - Tue, 25 Nov 2025 09:21:57 +0000 + MultiShotMaster: A Controllable Multi-Shot Video Generation Framework + https://arxiv.org/abs/2512.03041 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2512.03041.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Qinghe Wang, Xiaoyu Shi, Baolu Li, Weikang Bian, Quande Liu, Huchuan Lu, Xintao Wang, Pengfei Wan, Kun Gai, Xu Jia</p><p><b>Upvotes:</b> 57</p><p><b>Summary:</b> 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.</p> + https://arxiv.org/abs/2512.03041 + Tue, 02 Dec 2025 18:59:48 +0000 - UltraViCo: Breaking Extrapolation Limits in Video Diffusion Transformers - https://arxiv.org/abs/2511.20123 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.20123.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Min Zhao, Hongzhou Zhu, Yingze Wang, Bokai Yan, Jintao Zhang, Guande He, Ling Yang, Chongxuan Li, Jun Zhu</p><p><b>Upvotes:</b> 15</p><p><b>Summary:</b> Despite advances, video diffusion transformers still struggle to generalize beyond their training length, a challenge we term video length extrapolation. We identify two failure modes: model-specific periodic content repetition and a universal quality degradation. Prior works attempt to solve repetition via positional encodings, overlooking quality degradation and achieving only limited extrapolation. In this paper, we revisit this challenge from a more fundamental view: attention maps, which directly govern how context influences outputs. We identify that both failure modes arise from a unified cause: attention dispersion, where tokens beyond the training window dilute learned attention patterns. This leads to quality degradation and repetition emerges as a special case when this dispersion becomes structured into periodic attention patterns, induced by harmonic properties of positional encodings. Building on this insight, we propose UltraViCo, a training-free, plug-and-play method that suppresses attention for tokens beyond the training window via a constant decay factor. By jointly addressing both failure modes, we outperform a broad set of baselines largely across models and extrapolation ratios, pushing the extrapolation limit from 2x to 4x. Remarkably, it improves Dynamic Degree and Imaging Quality by 233% and 40.5% over the previous best method at 4x extrapolation. Furthermore, our method generalizes seamlessly to downstream tasks such as controllable video synthesis and editing.</p> - https://arxiv.org/abs/2511.20123 - Tue, 25 Nov 2025 09:44:10 +0000 + OneThinker: All-in-one Reasoning Model for Image and Video + https://arxiv.org/abs/2512.03043 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2512.03043.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Kaituo Feng, Manyuan Zhang, Hongyu Li, Kaixuan Fan, Shuang Chen, Yilei Jiang, Dian Zheng, Peiwen Sun, Yiyuan Zhang, Haoze Sun, Yan Feng, Peng Pei, Xunliang Cai, Xiangyu Yue</p><p><b>Upvotes:</b> 27</p><p><b>Summary:</b> 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.</p> + https://arxiv.org/abs/2512.03043 + Tue, 02 Dec 2025 18:59:52 +0000 - OmniAlpha: A Sequence-to-Sequence Framework for Unified Multi-Task RGBA Generation - https://arxiv.org/abs/2511.20211 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.20211.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Hao Yu, Jiabo Zhan, Zile Wang, Jinglin Wang, Huaisong Zhang, Hongyu Li, Xinrui Chen, Yongxian Wei, Chun Yuan</p><p><b>Upvotes:</b> 12</p><p><b>Summary:</b> Generative models have excelled in RGB synthesis, but real-world applications require RGBA manipulation. This has led to a fragmented landscape: specialized, single-task models handle alpha but lack versatility, while unified multi-task frameworks are confined to the RGB domain. To bridge this critical gap, we propose OmniAlpha, the first unified, multi-task generative framework for sequence-to-sequence RGBA image generation and editing. Its architecture features MSRoPE-BiL, a novel RoPE method with a bi-directionally extendable layer axis for its Diffusion Transformer (DiT) backbone, enabling the concurrent processing of multiple input and target RGBA layers. To power this framework, we introduce AlphaLayers, a new dataset of 1,000 high-quality, multi-layer triplets, built via a novel automated synthesis and filter pipeline. Jointly training OmniAlpha on this dataset across a comprehensive suite of 21 diverse tasks, extensive experiments demonstrate that our unified approach consistently outperforms strong, specialized baselines. Most notably, OmniAlpha achieves a dramatic 84.8% relative reduction in SAD for mask-free matting on AIM-500 and wins over 90% of human preferences in layer-conditioned completion. Our work proves that a unified, multi-task model can learn a superior shared representation for RGBA, paving the way for more powerful, layer-aware generative systems.</p> - https://arxiv.org/abs/2511.20211 - Tue, 25 Nov 2025 11:34:51 +0000 + MagicQuillV2: Precise and Interactive Image Editing with Layered Visual Cues + https://arxiv.org/abs/2512.03046 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2512.03046.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Zichen Liu, Yue Yu, Hao Ouyang, Qiuyu Wang, Shuailei Ma, Ka Leong Cheng, Wen Wang, Qingyan Bai, Yuxuan Zhang, Yanhong Zeng, Yixuan Li, Xing Zhu, Yujun Shen, Qifeng Chen</p><p><b>Upvotes:</b> 11</p><p><b>Summary:</b> We propose MagicQuill V2, a novel system that introduces a 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.</p> + https://arxiv.org/abs/2512.03046 + Tue, 02 Dec 2025 18:59:58 +0000 - Uplifting Table Tennis: A Robust, Real-World Application for 3D Trajectory and Spin Estimation - https://arxiv.org/abs/2511.20250 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.20250.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Daniel Kienzle, Katja Ludwig, Julian Lorenz, Shin'ichi Satoh, Rainer Lienhart</p><p><b>Upvotes:</b> 1</p><p><b>Summary:</b> Obtaining the precise 3D motion of a table tennis ball from standard monocular videos is a challenging problem, as existing methods trained on synthetic data struggle to generalize to the noisy, imperfect ball and table detections of the real world. This is primarily due to the inherent lack of 3D ground truth trajectories and spin annotations for real-world video. To overcome this, we propose a novel two-stage pipeline that divides the problem into a front-end perception task and a back-end 2D-to-3D uplifting task. This separation allows us to train the front-end components with abundant 2D supervision from our newly created TTHQ dataset, while the back-end uplifting network is trained exclusively on physically-correct synthetic data. We specifically re-engineer the uplifting model to be robust to common real-world artifacts, such as missing detections and varying frame rates. By integrating a ball detector and a table keypoint detector, our approach transforms a proof-of-concept uplifting method into a practical, robust, and high-performing end-to-end application for 3D table tennis trajectory and spin analysis.</p> - https://arxiv.org/abs/2511.20250 - Tue, 25 Nov 2025 12:25:20 +0000 + ViDiC: Video Difference Captioning + https://arxiv.org/abs/2512.03405 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2512.03405.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Jiangtao Wu, Shihao Li, Zhaozhou Bian, Yuanxing Zhang, Jialu Chen, Runzhe Wen, An Ping, Yiwen He, Jiakai Wang, Jiaheng Liu</p><p><b>Upvotes:</b> 24</p><p><b>Summary:</b> 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.</p> + https://arxiv.org/abs/2512.03405 + Wed, 03 Dec 2025 03:23:24 +0000 - The Image as Its Own Reward: Reinforcement Learning with Adversarial Reward for Image Generation - https://arxiv.org/abs/2511.20256 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.20256.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Weijia Mao, Hao Chen, Zhenheng Yang, Mike Zheng Shou</p><p><b>Upvotes:</b> 26</p><p><b>Summary:</b> A reliable reward function is essential for reinforcement learning (RL) in image generation. Most current RL approaches depend on pre-trained preference models that output scalar rewards to approximate human preferences. However, these rewards often fail to capture human perception and are vulnerable to reward hacking, where higher scores do not correspond to better images. To address this, we introduce Adv-GRPO, an RL framework with an adversarial reward that iteratively updates both the reward model and the generator. The reward model is supervised using reference images as positive samples and can largely avoid being hacked. Unlike KL regularization that constrains parameter updates, our learned reward directly guides the generator through its visual outputs, leading to higher-quality images. Moreover, while optimizing existing reward functions can alleviate reward hacking, their inherent biases remain. For instance, PickScore may degrade image quality, whereas OCR-based rewards often reduce aesthetic fidelity. To address this, we take the image itself as a reward, using reference images and vision foundation models (e.g., DINO) to provide rich visual rewards. These dense visual signals, instead of a single scalar, lead to consistent gains across image quality, aesthetics, and task-specific metrics. Finally, we show that combining reference samples with foundation-model rewards enables distribution transfer and flexible style customization. In human evaluation, our method outperforms Flow-GRPO and SD3, achieving 70.0% and 72.4% win rates in image quality and aesthetics, respectively. Code and models have been released.</p> - https://arxiv.org/abs/2511.20256 - Tue, 25 Nov 2025 12:35:57 +0000 + PretrainZero: Reinforcement Active Pretraining + https://arxiv.org/abs/2512.03442 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2512.03442.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Xingrun Xing, Zhiyuan Fan, Jie Lou, Guoqi Li, Jiajun Zhang, Debing Zhang</p><p><b>Upvotes:</b> 39</p><p><b>Summary:</b> Mimicking human behavior to actively learning from general experience and achieve artificial general intelligence has always been a human dream. Recent reinforcement learning (RL) based large-thinking models demonstrate impressive expert-level abilities, i.e., software and math, but still rely heavily on verifiable rewards in specific domains, placing a significant bottleneck to extend the performance boundary of general reasoning capabilities. In this work, we propose PretrainZero, a reinforcement active learning framework built on the pretraining corpus to extend RL from domain-specific post-training to general pretraining. PretrainZero features the following characteristics: 1) Active pretraining: inspired by the active learning ability of humans, PretrainZero learns a unified reasoning policy to actively identify reasonable and informative contents from pretraining corpus, and reason to predict these contents by RL. 2) Self-supervised learning: without any verifiable labels, pretrained reward models, or supervised fine-tuning, we directly pretrain reasoners from 3 to 30B base models on the general Wikipedia corpus using RL, significantly breaking the verification data-wall for general reasoning. 3) Verification scaling: by tackling increasingly challenging masked spans, PretrainZero substantially enhances the general reasoning abilities of pretrained base models. In reinforcement pretraining, PretrainZero improves Qwen3-4B-Base for 8.43, 5.96 and 10.60 on MMLU-Pro, SuperGPQA and math average benchmarks. In post-training, the pretrained models can also serve as reasoning foundation models for downstream RLVR tasks.</p> + https://arxiv.org/abs/2512.03442 + Wed, 03 Dec 2025 04:51:32 +0000 - Soft Adaptive Policy Optimization - https://arxiv.org/abs/2511.20347 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.20347.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Chang Gao, Chujie Zheng, Xiong-Hui Chen, Kai Dang, Shixuan Liu, Bowen Yu, An Yang, Shuai Bai, Jingren Zhou, Junyang Lin</p><p><b>Upvotes:</b> 31</p><p><b>Summary:</b> Reinforcement learning (RL) plays an increasingly important role in enhancing the reasoning capabilities of large language models (LLMs), yet stable and performant policy optimization remains challenging. Token-level importance ratios often exhibit high variance-a phenomenon exacerbated in Mixture-of-Experts models-leading to unstable updates. Existing group-based policy optimization methods, such as GSPO and GRPO, alleviate this problem via hard clipping, making it difficult to maintain both stability and effective learning. We propose Soft Adaptive Policy Optimization (SAPO), which replaces hard clipping with a smooth, temperature-controlled gate that adaptively attenuates off-policy updates while preserving useful learning signals. Compared with GSPO and GRPO, SAPO is both sequence-coherent and token-adaptive. Like GSPO, SAPO maintains sequence-level coherence, but its soft gating forms a continuous trust region that avoids the brittle hard clipping band used in GSPO. When a sequence contains a few highly off-policy tokens, GSPO suppresses all gradients for that sequence, whereas SAPO selectively down-weights only the offending tokens and preserves the learning signal from the near-on-policy ones, improving sample efficiency. Relative to GRPO, SAPO replaces hard token-level clipping with smooth, temperature-controlled scaling, enabling more informative and stable updates. Empirical results on mathematical reasoning benchmarks indicate that SAPO exhibits improved training stability and higher Pass@1 performance under comparable training budgets. Moreover, we employ SAPO to train the Qwen3-VL model series, demonstrating that SAPO yields consistent performance gains across diverse tasks and different model sizes. Overall, SAPO provides a more reliable, scalable, and effective optimization strategy for RL training of LLMs.</p> - https://arxiv.org/abs/2511.20347 - Tue, 25 Nov 2025 14:25:19 +0000 + Rethinking Prompt Design for Inference-time Scaling in Text-to-Visual Generation + https://arxiv.org/abs/2512.03534 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2512.03534.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Subin Kim, Sangwoo Mo, Mamshad Nayeem Rizve, Yiran Xu, Difan Liu, Jinwoo Shin, Tobias Hinz</p><p><b>Upvotes:</b> 18</p><p><b>Summary:</b> 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.</p> + https://arxiv.org/abs/2512.03534 + Wed, 03 Dec 2025 07:54:05 +0000 - Image-Free Timestep Distillation via Continuous-Time Consistency with Trajectory-Sampled Pairs - https://arxiv.org/abs/2511.20410 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.20410.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Bao Tang, Shuai Zhang, Yueting Zhu, Jijun Xiang, Xin Yang, Li Yu, Wenyu Liu, Xinggang Wang</p><p><b>Upvotes:</b> 2</p><p><b>Summary:</b> Timestep distillation is an effective approach for improving the generation efficiency of diffusion models. The Consistency Model (CM), as a trajectory-based framework, demonstrates significant potential due to its strong theoretical foundation and high-quality few-step generation. Nevertheless, current continuous-time consistency distillation methods still rely heavily on training data and computational resources, hindering their deployment in resource-constrained scenarios and limiting their scalability to diverse domains. To address this issue, we propose Trajectory-Backward Consistency Model (TBCM), which eliminates the dependence on external training data by extracting latent representations directly from the teacher model's generation trajectory. Unlike conventional methods that require VAE encoding and large-scale datasets, our self-contained distillation paradigm significantly improves both efficiency and simplicity. Moreover, the trajectory-extracted samples naturally bridge the distribution gap between training and inference, thereby enabling more effective knowledge transfer. Empirically, TBCM achieves 6.52 FID and 28.08 CLIP scores on MJHQ-30k under one-step generation, while reducing training time by approximately 40% compared to Sana-Sprint and saving a substantial amount of GPU memory, demonstrating superior efficiency without sacrificing quality. We further reveal the diffusion-generation space discrepancy in continuous-time consistency distillation and analyze how sampling strategies affect distillation performance, offering insights for future distillation research. GitHub Link: https://github.com/hustvl/TBCM.</p> - https://arxiv.org/abs/2511.20410 - Tue, 25 Nov 2025 15:36:20 +0000 + CookAnything: A Framework for Flexible and Consistent Multi-Step Recipe Image Generation + https://arxiv.org/abs/2512.03540 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2512.03540.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Ruoxuan Zhang, Bin Wen, Hongxia Xie, Yi Yao, Songhan Zuo, Jian-Yu Jiang-Lin, Hong-Han Shuai, Wen-Huang Cheng</p><p><b>Upvotes:</b> 11</p><p><b>Summary:</b> 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.</p> + https://arxiv.org/abs/2512.03540 + Wed, 03 Dec 2025 08:01:48 +0000 - MajutsuCity: Language-driven Aesthetic-adaptive City Generation with Controllable 3D Assets and Layouts - https://arxiv.org/abs/2511.20415 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.20415.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Zilong Huang, Jun He, Xiaobin Huang, Ziyi Xiong, Yang Luo, Junyan Ye, Weijia Li, Yiping Chen, Ting Han</p><p><b>Upvotes:</b> 8</p><p><b>Summary:</b> Generating realistic 3D cities is fundamental to world models, virtual reality, and game development, where an ideal urban scene must satisfy both stylistic diversity, fine-grained, and controllability. However, existing methods struggle to balance the creative flexibility offered by text-based generation with the object-level editability enabled by explicit structural representations. We introduce MajutsuCity, a natural language-driven and aesthetically adaptive framework for synthesizing structurally consistent and stylistically diverse 3D urban scenes. MajutsuCity represents a city as a composition of controllable layouts, assets, and materials, and operates through a four-stage pipeline. To extend controllability beyond initial generation, we further integrate MajutsuAgent, an interactive language-grounded editing agent} that supports five object-level operations. To support photorealistic and customizable scene synthesis, we also construct MajutsuDataset, a high-quality multimodal dataset} containing 2D semantic layouts and height maps, diverse 3D building assets, and curated PBR materials and skyboxes, each accompanied by detailed annotations. Meanwhile, we develop a practical set of evaluation metrics, covering key dimensions such as structural consistency, scene complexity, material fidelity, and lighting atmosphere. Extensive experiments demonstrate MajutsuCity reduces layout FID by 83.7% compared with CityDreamer and by 20.1% over CityCraft. Our method ranks first across all AQS and RDR scores, outperforming existing methods by a clear margin. These results confirm MajutsuCity as a new state-of-the-art in geometric fidelity, stylistic adaptability, and semantic controllability for 3D city generation. We expect our framework can inspire new avenues of research in 3D city generation. Our dataset and code will be released at https://github.com/LongHZ140516/MajutsuCity.</p> - https://arxiv.org/abs/2511.20415 - Tue, 25 Nov 2025 15:40:12 +0000 + Thinking with Programming Vision: Towards a Unified View for Thinking with Images + https://arxiv.org/abs/2512.03746 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2512.03746.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Zirun Guo, Minjie Hong, Feng Zhang, Kai Jia, Tao Jin</p><p><b>Upvotes:</b> 15</p><p><b>Summary:</b> 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.</p> + https://arxiv.org/abs/2512.03746 + Wed, 03 Dec 2025 12:44:15 +0000 - Block Cascading: Training Free Acceleration of Block-Causal Video Models - https://arxiv.org/abs/2511.20426 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.20426.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Hmrishav Bandyopadhyay, Nikhil Pinnaparaju, Rahim Entezari, Jim Scott, Yi-Zhe Song, Varun Jampani</p><p><b>Upvotes:</b> 7</p><p><b>Summary:</b> Block-causal video generation faces a stark speed-quality trade-off: small 1.3B models manage only 16 FPS while large 14B models crawl at 4.5 FPS, forcing users to choose between responsiveness and quality. Block Cascading significantly mitigates this trade-off through training-free parallelization. Our key insight: future video blocks do not need fully denoised current blocks to begin generation. By starting block generation with partially denoised context from predecessors, we transform sequential pipelines into parallel cascades where multiple blocks denoise simultaneously. With 5 GPUs exploiting temporal parallelism, we achieve ~2x acceleration across all model scales: 1.3B models accelerate from 16 to 30 FPS, 14B models from 4.5 to 12.5 FPS. Beyond inference speed, Block Cascading eliminates overhead from KV-recaching (of ~200ms) during context switches for interactive generation. Extensive evaluations validated against multiple block-causal pipelines demonstrate no significant loss in generation quality when switching from block-causal to Block Cascading pipelines for inference. Project Page: https://hmrishavbandy.github.io/block_cascading_page/</p> - https://arxiv.org/abs/2511.20426 - Tue, 25 Nov 2025 15:52:58 +0000 + Jina-VLM: Small Multilingual Vision Language Model + https://arxiv.org/abs/2512.04032 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2512.04032.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Andreas Koukounas, Georgios Mastrapas, Florian Hönicke, Sedigheh Eslami, Guillaume Roncari, Scott Martens, Han Xiao</p><p><b>Upvotes:</b> 9</p><p><b>Summary:</b> We present Jina-VLM, a 2.4B parameter vision-language model that achieves state-of-the-art multilingual visual question answering among open 2B-scale VLMs. The model couples a SigLIP2 vision encoder with a Qwen3 language backbone through an attention-pooling connector that enables token-efficient processing of arbitrary-resolution images. Across standard VQA benchmarks and multilingual evaluations, Jina-VLM outperforms comparable models while preserving competitive text-only performance.</p> + https://arxiv.org/abs/2512.04032 + Wed, 03 Dec 2025 18:13:41 +0000 - STARFlow-V: End-to-End Video Generative Modeling with Normalizing Flow - https://arxiv.org/abs/2511.20462 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.20462.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Jiatao Gu, Ying Shen, Tianrong Chen, Laurent Dinh, Yuyang Wang, Miguel Angel Bautista, David Berthelot, Josh Susskind, Shuangfei Zhai</p><p><b>Upvotes:</b> 17</p><p><b>Summary:</b> Normalizing flows (NFs) are end-to-end likelihood-based generative models for continuous data, and have recently regained attention with encouraging progress on image generation. Yet in the video generation domain, where spatiotemporal complexity and computational cost are substantially higher, state-of-the-art systems almost exclusively rely on diffusion-based models. In this work, we revisit this design space by presenting STARFlow-V, a normalizing flow-based video generator with substantial benefits such as end-to-end learning, robust causal prediction, and native likelihood estimation. Building upon the recently proposed STARFlow, STARFlow-V operates in the spatiotemporal latent space with a global-local architecture which restricts causal dependencies to a global latent space while preserving rich local within-frame interactions. This eases error accumulation over time, a common pitfall of standard autoregressive diffusion model generation. Additionally, we propose flow-score matching, which equips the model with a light-weight causal denoiser to improve the video generation consistency in an autoregressive fashion. To improve the sampling efficiency, STARFlow-V employs a video-aware Jacobi iteration scheme that recasts inner updates as parallelizable iterations without breaking causality. Thanks to the invertible structure, the same model can natively support text-to-video, image-to-video as well as video-to-video generation tasks. Empirically, STARFlow-V achieves strong visual fidelity and temporal consistency with practical sampling throughput relative to diffusion-based baselines. These results present the first evidence, to our knowledge, that NFs are capable of high-quality autoregressive video generation, establishing them as a promising research direction for building world models. Code and generated samples are available at https://github.com/apple/ml-starflow.</p> - https://arxiv.org/abs/2511.20462 - Tue, 25 Nov 2025 16:27:58 +0000 + RELIC: Interactive Video World Model with Long-Horizon Memory + https://arxiv.org/abs/2512.04040 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2512.04040.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Yicong Hong, Yiqun Mei, Chongjian Ge, Yiran Xu, Yang Zhou, Sai Bi, Yannick Hold-Geoffroy, Mike Roberts, Matthew Fisher, Eli Shechtman, Kalyan Sunkavalli, Feng Liu, Zhengqi Li, Hao Tan</p><p><b>Upvotes:</b> 19</p><p><b>Summary:</b> 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.</p> + https://arxiv.org/abs/2512.04040 + Wed, 03 Dec 2025 18:29:20 +0000 - NVIDIA Nemotron Parse 1.1 - https://arxiv.org/abs/2511.20478 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.20478.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Kateryna Chumachenko, Amala Sanjay Deshmukh, Jarno Seppanen, Ilia Karmanov, Chia-Chih Chen, Lukas Voegtle, Philipp Fischer, Marek Wawrzos, Saeid Motiian, Roman Ageev, Kedi Wu, Alexandre Milesi, Maryam Moosaei, Krzysztof Pawelec, Padmavathy Subramanian, Mehrzad Samadi, Xin Yu, Celina Dear, Sarah Stoddard, Jenna Diamond, Jesse Oliver, Leanna Chraghchian, Patrick Skelly, Tom Balough, Yao Xu, Jane Polak Scowcroft, Daniel Korzekwa, Darragh Hanley, Sandip Bhaskar, Timo Roman, Karan Sapra, Andrew Tao, Bryan Catanzaro</p><p><b>Upvotes:</b> 18</p><p><b>Summary:</b> We introduce Nemotron-Parse-1.1, a lightweight document parsing and OCR model that advances the capabilities of its predecessor, Nemoretriever-Parse-1.0. Nemotron-Parse-1.1 delivers improved capabilities across general OCR, markdown formatting, structured table parsing, and text extraction from pictures, charts, and diagrams. It also supports a longer output sequence length for visually dense documents. As with its predecessor, it extracts bounding boxes of text segments, as well as corresponding semantic classes. Nemotron-Parse-1.1 follows an encoder-decoder architecture with 885M parameters, including a compact 256M-parameter language decoder. It achieves competitive accuracy on public benchmarks making it a strong lightweight OCR solution. We release the model weights publicly on Huggingface, as well as an optimized NIM container, along with a subset of the training data as part of the broader Nemotron-VLM-v2 dataset. Additionally, we release Nemotron-Parse-1.1-TC which operates on a reduced vision token length, offering a 20% speed improvement with minimal quality degradation.</p> - https://arxiv.org/abs/2511.20478 - Tue, 25 Nov 2025 16:41:25 +0000 + SpaceTools: Tool-Augmented Spatial Reasoning via Double Interactive RL + https://arxiv.org/abs/2512.04069 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2512.04069.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Siyi Chen, Mikaela Angelina Uy, Chan Hee Song, Faisal Ladhak, Adithyavairavan Murali, Qing Qu, Stan Birchfield, Valts Blukis, Jonathan Tremblay</p><p><b>Upvotes:</b> 21</p><p><b>Summary:</b> 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/.</p> + https://arxiv.org/abs/2512.04069 + Wed, 03 Dec 2025 18:50:04 +0000 - Does Understanding Inform Generation in Unified Multimodal Models? From Analysis to Path Forward - https://arxiv.org/abs/2511.20561 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.20561.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Yuwei Niu, Weiyang Jin, Jiaqi Liao, Chaoran Feng, Peng Jin, Bin Lin, Zongjian Li, Bin Zhu, Weihao Yu, Li Yuan</p><p><b>Upvotes:</b> 31</p><p><b>Summary:</b> Recent years have witnessed significant progress in Unified Multimodal Models, yet a fundamental question remains: Does understanding truly inform generation? To investigate this, we introduce UniSandbox, a decoupled evaluation framework paired with controlled, synthetic datasets to avoid data leakage and enable detailed analysis. Our findings reveal a significant understanding-generation gap, which is mainly reflected in two key dimensions: reasoning generation and knowledge transfer. Specifically, for reasoning generation tasks, we observe that explicit Chain-of-Thought (CoT) in the understanding module effectively bridges the gap, and further demonstrate that a self-training approach can successfully internalize this ability, enabling implicit reasoning during generation. Additionally, for knowledge transfer tasks, we find that CoT assists the generative process by helping retrieve newly learned knowledge, and also discover that query-based architectures inherently exhibit latent CoT-like properties that affect this transfer. UniSandbox provides preliminary insights for designing future unified architectures and training strategies that truly bridge the gap between understanding and generation. Code and data are available at https://github.com/PKU-YuanGroup/UniSandBox</p> - https://arxiv.org/abs/2511.20561 - Tue, 25 Nov 2025 17:58:48 +0000 + PosterCopilot: Toward Layout Reasoning and Controllable Editing for Professional Graphic Design + https://arxiv.org/abs/2512.04082 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2512.04082.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Jiazhe Wei, Ken Li, Tianyu Lao, Haofan Wang, Liang Wang, Caifeng Shan, Chenyang Si</p><p><b>Upvotes:</b> 11</p><p><b>Summary:</b> Graphic design forms the cornerstone of modern visual communication, serving as a vital medium for promoting cultural and commercial events. Recent advances have explored automating this process using Large Multimodal Models (LMMs), yet existing methods often produce geometrically inaccurate layouts and lack the iterative, layer-specific editing required in professional workflows. To address these limitations, we present PosterCopilot, a framework that advances layout reasoning and controllable editing for professional graphic design. Specifically, we introduce a progressive three-stage training strategy that equips LMMs with geometric understanding and aesthetic reasoning for layout design, consisting of Perturbed Supervised Fine-Tuning, Reinforcement Learning for Visual-Reality Alignment, and Reinforcement Learning from Aesthetic Feedback. Furthermore, we develop a complete workflow that couples the trained LMM-based design model with generative models, enabling layer-controllable, iterative editing for precise element refinement while maintaining global visual consistency. Extensive experiments demonstrate that PosterCopilot achieves geometrically accurate and aesthetically superior layouts, offering unprecedented controllability for professional iterative design.</p> + https://arxiv.org/abs/2512.04082 + Wed, 03 Dec 2025 18:59:37 +0000 - PhysChoreo: Physics-Controllable Video Generation with Part-Aware Semantic Grounding - https://arxiv.org/abs/2511.20562 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.20562.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Haoze Zhang, Tianyu Huang, Zichen Wan, Xiaowei Jin, Hongzhi Zhang, Hui Li, Wangmeng Zuo</p><p><b>Upvotes:</b> 4</p><p><b>Summary:</b> While recent video generation models have achieved significant visual fidelity, they often suffer from the lack of explicit physical controllability and plausibility. To address this, some recent studies attempted to guide the video generation with physics-based rendering. However, these methods face inherent challenges in accurately modeling complex physical properties and effectively control ling the resulting physical behavior over extended temporal sequences. In this work, we introduce PhysChoreo, a novel framework that can generate videos with diverse controllability and physical realism from a single image. Our method consists of two stages: first, it estimates the static initial physical properties of all objects in the image through part-aware physical property reconstruction. Then, through temporally instructed and physically editable simulation, it synthesizes high-quality videos with rich dynamic behaviors and physical realism. Experimental results show that PhysChoreo can generate videos with rich behaviors and physical realism, outperforming state-of-the-art methods on multiple evaluation metrics.</p> - https://arxiv.org/abs/2511.20562 - Tue, 25 Nov 2025 17:59:04 +0000 + On GRPO Collapse in Search-R1: The Lazy Likelihood-Displacement Death Spiral + https://arxiv.org/abs/2512.04220 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2512.04220.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Wenlong Deng, Yushu Li, Boying Gong, Yi Ren, Christos Thrampoulidis, Xiaoxiao Li</p><p><b>Upvotes:</b> 9</p><p><b>Summary:</b> 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.</p> + https://arxiv.org/abs/2512.04220 + Wed, 03 Dec 2025 19:41:15 +0000 - VQ-VA World: Towards High-Quality Visual Question-Visual Answering - https://arxiv.org/abs/2511.20573 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.20573.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Chenhui Gou, Zilong Chen, Zeyu Wang, Feng Li, Deyao Zhu, Zicheng Duan, Kunchang Li, Chaorui Deng, Hongyi Yuan, Haoqi Fan, Cihang Xie, Jianfei Cai, Hamid Rezatofighi</p><p><b>Upvotes:</b> 7</p><p><b>Summary:</b> This paper studies Visual Question-Visual Answering (VQ-VA): generating an image, rather than text, in response to a visual question -- an ability that has recently emerged in proprietary systems such as NanoBanana and GPT-Image. To also bring this capability to open-source models, we introduce VQ-VA World, a data-centric framework built around an agentic pipeline for large-scale, targeted data construction. Leveraging web-scale deployment, this pipeline crawls a massive amount of ~1.8M high-quality, interleaved image-text samples for model training. For evaluation, we further release IntelligentBench, a human-curated benchmark that systematically assesses VQ-VA along the aspects of world knowledge, design knowledge, and reasoning. Training with VQ-VA World data yields strong empirical gains: it helps LightFusion attain 53.06 on IntelligentBench, substantially surpassing the best prior open-source baselines (i.e., 7.78 from vanilla LightFusion; 1.94 from UniWorld-V1), and significantly narrowing the gap toward leading proprietary systems (e.g., 81.67 from NanoBanana; 82.64 from GPT-Image). By releasing the full suite of model weights, datasets, and pipelines, we hope to stimulate future research on VQ-VA.</p> - https://arxiv.org/abs/2511.20573 - Tue, 25 Nov 2025 18:06:22 +0000 + DAComp: Benchmarking Data Agents across the Full Data Intelligence Lifecycle + https://arxiv.org/abs/2512.04324 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2512.04324.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Fangyu Lei, Jinxiang Meng, Yiming Huang, Junjie Zhao, Yitong Zhang, Jianwen Luo, Xin Zou, Ruiyi Yang, Wenbo Shi, Yan Gao, Shizhu He, Zuo Wang, Qian Liu, Yang Wang, Ke Wang, Jun Zhao, Kang Liu</p><p><b>Upvotes:</b> 128</p><p><b>Summary:</b> 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</p> + https://arxiv.org/abs/2512.04324 + Wed, 03 Dec 2025 23:21:28 +0000 - ROOT: Robust Orthogonalized Optimizer for Neural Network Training - https://arxiv.org/abs/2511.20626 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.20626.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Wei He, Kai Han, Hang Zhou, Hanting Chen, Zhicheng Liu, Xinghao Chen, Yunhe Wang</p><p><b>Upvotes:</b> 164</p><p><b>Summary:</b> The optimization of large language models (LLMs) remains a critical challenge, particularly as model scaling exacerbates sensitivity to algorithmic imprecision and training instability. Recent advances in optimizers have improved convergence efficiency through momentum orthogonalization, but suffer from two key robustness limitations: dimensional fragility in orthogonalization precision and vulnerability to outlier-induced noise. To address these robustness challenges, we introduce ROOT, a Robust Orthogonalized Optimizer that enhances training stability through dual robustness mechanisms. First, we develop a dimension-robust orthogonalization scheme using adaptive Newton iterations with fine-grained coefficients tailored to specific matrix sizes, ensuring consistent precision across diverse architectural configurations. Second, we introduce an optimization-robust framework via proximal optimization that suppresses outlier noise while preserving meaningful gradient directions. Extensive experiments demonstrate that ROOT achieves significantly improved robustness, with faster convergence and superior final performance compared to both Muon and Adam-based optimizers, particularly in noisy and non-convex scenarios. Our work establishes a new paradigm for developing robust and precise optimizers capable of handling the complexities of modern large-scale model training. The code will be available at https://github.com/huawei-noah/noah-research/tree/master/ROOT.</p> - https://arxiv.org/abs/2511.20626 - Tue, 25 Nov 2025 18:48:05 +0000 + UltraImage: Rethinking Resolution Extrapolation in Image Diffusion Transformers + https://arxiv.org/abs/2512.04504 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2512.04504.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Min Zhao, Bokai Yan, Xue Yang, Hongzhou Zhu, Jintao Zhang, Shilong Liu, Chongxuan Li, Jun Zhu</p><p><b>Upvotes:</b> 13</p><p><b>Summary:</b> 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 https://thu-ml.github.io/ultraimage.github.io/{https://thu-ml.github.io/ultraimage.github.io/}.</p> + https://arxiv.org/abs/2512.04504 + Thu, 04 Dec 2025 06:24:04 +0000 - Reinforcing Action Policies by Prophesying - https://arxiv.org/abs/2511.20633 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.20633.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Jiahui Zhang, Ze Huang, Chun Gu, Zipei Ma, Li Zhang</p><p><b>Upvotes:</b> 2</p><p><b>Summary:</b> Vision-Language-Action (VLA) policies excel in aligning language, perception, and robot control. However, most VLAs are trained purely by imitation, which overfits to demonstrations, and is brittle under distribution shift. Reinforcement learning (RL) directly optimizes task reward and thus addresses this misalignment, but real-robot interaction is expensive and conventional simulators are hard to engineer and transfer. We address both data efficiency and optimization stability in VLA post-training via a learned world model and an RL procedure tailored to flow-based action heads. Specifically, we introduce Prophet, a unified action-to-video robot actuation pretrained across large-scale, heterogeneous robot data to learn reusable action-outcome dynamics. It is able to few-shot adapt to new robots, objects, and environments, yielding a rollout-ready simulator. Upon Prophet, we reinforce action policies with Flow-action-GRPO (FA-GRPO), which adapts Flow-GRPO to operate on VLA actions, and with FlowScale, a stepwise reweighting that rescales per-step gradients in the flow head. Together, Prophet, FA-GRPO, and FlowScale constitute ProphRL, a practical, data- and compute-efficient path to VLA post-training. Experiments show 5-17% success gains on public benchmarks and 24-30% gains on real robots across different VLA variants.</p> - https://arxiv.org/abs/2511.20633 - Tue, 25 Nov 2025 18:52:56 +0000 + Live Avatar: Streaming Real-time Audio-Driven Avatar Generation with Infinite Length + https://arxiv.org/abs/2512.04677 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2512.04677.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Yubo Huang, Hailong Guo, Fangtai Wu, Shifeng Zhang, Shijie Huang, Qijun Gan, Lin Liu, Sirui Zhao, Enhong Chen, Jiaming Liu, Steven Hoi</p><p><b>Upvotes:</b> 139</p><p><b>Summary:</b> 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.</p> + https://arxiv.org/abs/2512.04677 + Thu, 04 Dec 2025 11:11:24 +0000 - iMontage: Unified, Versatile, Highly Dynamic Many-to-many Image Generation - https://arxiv.org/abs/2511.20635 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.20635.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Zhoujie Fu, Xianfang Zeng, Jinghong Lan, Xinyao Liao, Cheng Chen, Junyi Chen, Jiacheng Wei, Wei Cheng, Shiyu Liu, Yunuo Chen, Gang Yu, Guosheng Lin</p><p><b>Upvotes:</b> 30</p><p><b>Summary:</b> Pre-trained video models learn powerful priors for generating high-quality, temporally coherent content. While these models excel at temporal coherence, their dynamics are often constrained by the continuous nature of their training data. We hypothesize that by injecting the rich and unconstrained content diversity from image data into this coherent temporal framework, we can generate image sets that feature both natural transitions and a far more expansive dynamic range. To this end, we introduce iMontage, a unified framework designed to repurpose a powerful video model into an all-in-one image generator. The framework consumes and produces variable-length image sets, unifying a wide array of image generation and editing tasks. To achieve this, we propose an elegant and minimally invasive adaptation strategy, complemented by a tailored data curation process and training paradigm. This approach allows the model to acquire broad image manipulation capabilities without corrupting its invaluable original motion priors. iMontage excels across several mainstream many-in-many-out tasks, not only maintaining strong cross-image contextual consistency but also generating scenes with extraordinary dynamics that surpass conventional scopes. Find our homepage at: https://kr1sjfu.github.io/iMontage-web/.</p> - https://arxiv.org/abs/2511.20635 - Tue, 25 Nov 2025 18:54:16 +0000 + Reward Forcing: Efficient Streaming Video Generation with Rewarded Distribution Matching Distillation + https://arxiv.org/abs/2512.04678 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2512.04678.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Yunhong Lu, Yanhong Zeng, Haobo Li, Hao Ouyang, Qiuyu Wang, Ka Leong Cheng, Jiapeng Zhu, Hengyuan Cao, Zhipeng Zhang, Xing Zhu, Yujun Shen, Min Zhang</p><p><b>Upvotes:</b> 33</p><p><b>Summary:</b> 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.</p> + https://arxiv.org/abs/2512.04678 + Thu, 04 Dec 2025 11:12:13 +0000 - Latent Collaboration in Multi-Agent Systems - https://arxiv.org/abs/2511.20639 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.20639.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Jiaru Zou, Xiyuan Yang, Ruizhong Qiu, Gaotang Li, Katherine Tieu, Pan Lu, Ke Shen, Hanghang Tong, Yejin Choi, Jingrui He, James Zou, Mengdi Wang, Ling Yang</p><p><b>Upvotes:</b> 89</p><p><b>Summary:</b> Multi-agent systems (MAS) extend large language models (LLMs) from independent single-model reasoning to coordinative system-level intelligence. While existing LLM agents depend on text-based mediation for reasoning and communication, we take a step forward by enabling models to collaborate directly within the continuous latent space. We introduce LatentMAS, an end-to-end training-free framework that enables pure latent collaboration among LLM agents. In LatentMAS, each agent first performs auto-regressive latent thoughts generation through last-layer hidden embeddings. A shared latent working memory then preserves and transfers each agent's internal representations, ensuring lossless information exchange. We provide theoretical analyses establishing that LatentMAS attains higher expressiveness and lossless information preservation with substantially lower complexity than vanilla text-based MAS. In addition, empirical evaluations across 9 comprehensive benchmarks spanning math and science reasoning, commonsense understanding, and code generation show that LatentMAS consistently outperforms strong single-model and text-based MAS baselines, achieving up to 14.6% higher accuracy, reducing output token usage by 70.8%-83.7%, and providing 4x-4.3x faster end-to-end inference. These results demonstrate that our new latent collaboration framework enhances system-level reasoning quality while offering substantial efficiency gains without any additional training. Code and data are fully open-sourced at https://github.com/Gen-Verse/LatentMAS.</p> - https://arxiv.org/abs/2511.20639 - Tue, 25 Nov 2025 18:56:57 +0000 + SignRoundV2: Closing the Performance Gap in Extremely Low-Bit Post-Training Quantization for LLMs + https://arxiv.org/abs/2512.04746 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2512.04746.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Wenhua Cheng, Weiwei Zhang, Heng Guo, Haihao Shen</p><p><b>Upvotes:</b> 10</p><p><b>Summary:</b> 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.</p> + https://arxiv.org/abs/2512.04746 + Thu, 04 Dec 2025 12:35:10 +0000 - Concept-Aware Batch Sampling Improves Language-Image Pretraining - https://arxiv.org/abs/2511.20643 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.20643.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Adhiraj Ghosh, Vishaal Udandarao, Thao Nguyen, Matteo Farina, Mehdi Cherti, Jenia Jitsev, Sewoong Oh, Elisa Ricci, Ludwig Schmidt, Matthias Bethge</p><p><b>Upvotes:</b> 1</p><p><b>Summary:</b> What data should a vision-language model be trained on? To answer this question, many data curation efforts center on the quality of a dataset. However, most of these existing methods are (i) offline, i.e. they produce a static dataset from a set of predetermined filtering criteria, and (ii) concept-agnostic, i.e. they use model-based filters which induce additional data biases. In this work, we go beyond such offline, concept-agnostic methods and advocate for more flexible, task-adaptive online concept-based curation. Our first contribution is DataConcept, a collection of 128M web-crawled image-text pairs annotated with fine-grained details about their concept composition. Building on DataConcept, we introduce Concept-Aware Batch Sampling (CABS), a simple yet effective batch sampling framework that flexibly constructs batches on-the-fly based on specific target distributions. We propose two variants: (i) Diversity Maximization (CABS-DM) to curate batches with a broad coverage of available concepts, and (ii) Frequency Maximization (CABS-FM) to curate batches with high object multiplicity. Through extensive evaluations across 28 benchmarks, we demonstrate that our CABS method significantly benefits CLIP/SigLIP model classes and yields highly performant models. Overall, CABS represents a strong open-source alternative to proprietary online data curation algorithms, enabling practitioners to define custom concept distributions that optimize for specific downstream tasks.</p> - https://arxiv.org/abs/2511.20643 - Tue, 25 Nov 2025 18:58:07 +0000 + SIMA 2: A Generalist Embodied Agent for Virtual Worlds + https://arxiv.org/abs/2512.04797 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2512.04797.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> SIMA team, Adrian Bolton, Alexander Lerchner, Alexandra Cordell, Alexandre Moufarek, Andrew Bolt, Andrew Lampinen, Anna Mitenkova, Arne Olav Hallingstad, Bojan Vujatovic, Bonnie Li, Cong Lu, Daan Wierstra, Daniel P. Sawyer, Daniel Slater, David Reichert, Davide Vercelli, Demis Hassabis, Drew A. Hudson, Duncan Williams, Ed Hirst, Fabio Pardo, Felix Hill, Frederic Besse, Hannah Openshaw, Harris Chan, Hubert Soyer, Jane X. Wang, Jeff Clune, John Agapiou, John Reid, Joseph Marino, Junkyung Kim, Karol Gregor, Kaustubh Sridhar, Kay McKinney, Laura Kampis, Lei M. Zhang, Loic Matthey, Luyu Wang, Maria Abi Raad, Maria Loks-Thompson, Martin Engelcke, Matija Kecman, Matthew Jackson, Maxime Gazeau, Ollie Purkiss, Oscar Knagg, Peter Stys, Piermaria Mendolicchio, Raia Hadsell, Rosemary Ke, Ryan Faulkner, Sarah Chakera, Satinder Singh Baveja, Shane Legg, Sheleem Kashem, Tayfun Terzi, Thomas Keck, Tim Harley, Tim Scholtes, Tyson Roberts, Volodymyr Mnih, Yulan Liu, Zhengdong Wang, Zoubin Ghahramani</p><p><b>Upvotes:</b> 11</p><p><b>Summary:</b> We introduce SIMA 2, a generalist embodied agent that understands and acts in a wide variety of 3D virtual worlds. Built upon a Gemini foundation model, SIMA 2 represents a significant step toward active, goal-directed interaction within an embodied environment. Unlike prior work (e.g., SIMA 1) limited to simple language commands, SIMA 2 acts as an interactive partner, capable of reasoning about high-level goals, conversing with the user, and handling complex instructions given through language and images. Across a diverse portfolio of games, SIMA 2 substantially closes the gap with human performance and demonstrates robust generalization to previously unseen environments, all while retaining the base model's core reasoning capabilities. Furthermore, we demonstrate a capacity for open-ended self-improvement: by leveraging Gemini to generate tasks and provide rewards, SIMA 2 can autonomously learn new skills from scratch in a new environment. This work validates a path toward creating versatile and continuously learning agents for both virtual and, eventually, physical worlds.</p> + https://arxiv.org/abs/2512.04797 + Thu, 04 Dec 2025 13:46:11 +0000 - Diverse Video Generation with Determinantal Point Process-Guided Policy Optimization - https://arxiv.org/abs/2511.20647 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.20647.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Tahira Kazimi, Connor Dunlop, Pinar Yanardag</p><p><b>Upvotes:</b> 2</p><p><b>Summary:</b> While recent text-to-video (T2V) diffusion models have achieved impressive quality and prompt alignment, they often produce low-diversity outputs when sampling multiple videos from a single text prompt. We tackle this challenge by formulating it as a set-level policy optimization problem, with the goal of training a policy that can cover the diverse range of plausible outcomes for a given prompt. To address this, we introduce DPP-GRPO, a novel framework for diverse video generation that combines Determinantal Point Processes (DPPs) and Group Relative Policy Optimization (GRPO) theories to enforce explicit reward on diverse generations. Our objective turns diversity into an explicit signal by imposing diminishing returns on redundant samples (via DPP) while supplies groupwise feedback over candidate sets (via GRPO). Our framework is plug-and-play and model-agnostic, and encourages diverse generations across visual appearance, camera motions, and scene structure without sacrificing prompt fidelity or perceptual quality. We implement our method on WAN and CogVideoX, and show that our method consistently improves video diversity on state-of-the-art benchmarks such as VBench, VideoScore, and human preference studies. Moreover, we release our code and a new benchmark dataset of 30,000 diverse prompts to support future research.</p> - https://arxiv.org/abs/2511.20647 - Tue, 25 Nov 2025 18:59:45 +0000 + Model-Based and Sample-Efficient AI-Assisted Math Discovery in Sphere Packing + https://arxiv.org/abs/2512.04829 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2512.04829.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Rasul Tutunov, Alexandre Maraval, Antoine Grosnit, Xihan Li, Jun Wang, Haitham Bou-Ammar</p><p><b>Upvotes:</b> 10</p><p><b>Summary:</b> 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.</p> + https://arxiv.org/abs/2512.04829 + Thu, 04 Dec 2025 14:11:52 +0000 - SPHINX: A Synthetic Environment for Visual Perception and Reasoning - https://arxiv.org/abs/2511.20814 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.20814.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Md Tanvirul Alam, Saksham Aggarwal, Justin Yang Chae, Nidhi Rastogi</p><p><b>Upvotes:</b> 2</p><p><b>Summary:</b> We present Sphinx, a synthetic environment for visual perception and reasoning that targets core cognitive primitives. Sphinx procedurally generates puzzles using motifs, tiles, charts, icons, and geometric primitives, each paired with verifiable ground-truth solutions, enabling both precise evaluation and large-scale dataset construction. The benchmark covers 25 task types spanning symmetry detection, geometric transformations, spatial reasoning, chart interpretation, and sequence prediction. Evaluating recent large vision-language models (LVLMs) shows that even state-of-the-art GPT-5 attains only 51.1% accuracy, well below human performance. Finally, we demonstrate that reinforcement learning with verifiable rewards (RLVR) substantially improves model accuracy on these tasks and yields gains on external visual reasoning benchmarks, highlighting its promise for advancing multimodal reasoning.</p> - https://arxiv.org/abs/2511.20814 - Tue, 25 Nov 2025 20:00:47 +0000 + Semantics Lead the Way: Harmonizing Semantic and Texture Modeling with Asynchronous Latent Diffusion + https://arxiv.org/abs/2512.04926 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2512.04926.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Yueming Pan, Ruoyu Feng, Qi Dai, Yuqi Wang, Wenfeng Lin, Mingyu Guo, Chong Luo, Nanning Zheng</p><p><b>Upvotes:</b> 32</p><p><b>Summary:</b> 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/.</p> + https://arxiv.org/abs/2512.04926 + Thu, 04 Dec 2025 15:57:27 +0000 - ENACT: Evaluating Embodied Cognition with World Modeling of Egocentric Interaction - https://arxiv.org/abs/2511.20937 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.20937.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Qineng Wang, Wenlong Huang, Yu Zhou, Hang Yin, Tianwei Bao, Jianwen Lyu, Weiyu Liu, Ruohan Zhang, Jiajun Wu, Li Fei-Fei, Manling Li</p><p><b>Upvotes:</b> 11</p><p><b>Summary:</b> Embodied cognition argues that intelligence arises from sensorimotor interaction rather than passive observation. It raises an intriguing question: do modern vision-language models (VLMs), trained largely in a disembodied manner, exhibit signs of embodied cognition? We introduce ENACT, a benchmark that casts evaluation of embodied cognition as world modeling from egocentric interaction in a visual question answering (VQA) format. Framed as a partially observable Markov decision process (POMDP) whose actions are scene graph changes, ENACT comprises two complementary sequence reordering tasks: forward world modeling (reorder shuffled observations given actions) and inverse world modeling (reorder shuffled actions given observations). While conceptually simple, solving these tasks implicitly demands capabilities central to embodied cognition-affordance recognition, action-effect reasoning, embodied awareness, and interactive, long-horizon memory from partially observable egocentric input, while avoiding low-level image synthesis that could confound the evaluation. We provide a scalable pipeline that synthesizes QA pairs from robotics simulation (BEHAVIOR) and evaluates models on 8,972 QA pairs spanning long-horizon home-scale activities. Experiments reveal a performance gap between frontier VLMs and humans that widens with interaction horizon. Models consistently perform better on the inverse task than the forward one and exhibit anthropocentric biases, including a preference for right-handed actions and degradation when camera intrinsics or viewpoints deviate from human vision. Website at https://enact-embodied-cognition.github.io/.</p> - https://arxiv.org/abs/2511.20937 - Wed, 26 Nov 2025 00:06:02 +0000 + Aligned but Stereotypical? The Hidden Influence of System Prompts on Social Bias in LVLM-Based Text-to-Image Models + https://arxiv.org/abs/2512.04981 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2512.04981.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> NaHyeon Park, Namin An, Kunhee Kim, Soyeon Yoon, Jiahao Huo, Hyunjung Shim</p><p><b>Upvotes:</b> 6</p><p><b>Summary:</b> Large vision-language model (LVLM) based text-to-image (T2I) systems have become the dominant paradigm in image generation, yet whether they amplify social biases remains insufficiently understood. In this paper, we show that LVLM-based models produce markedly more socially biased images than non-LVLM-based models. We introduce a 1,024 prompt benchmark spanning four levels of linguistic complexity and evaluate demographic bias across multiple attributes in a systematic manner. Our analysis identifies system prompts, the predefined instructions guiding LVLMs, as a primary driver of biased behavior. Through decoded intermediate representations, token-probability diagnostics, and embedding-association analyses, we reveal how system prompts encode demographic priors that propagate into image synthesis. To this end, we propose FairPro, a training-free meta-prompting framework that enables LVLMs to self-audit and construct fairness-aware system prompts at test time. Experiments on two LVLM-based T2I models, SANA and Qwen-Image, show that FairPro substantially reduces demographic bias while preserving text-image alignment. We believe our findings provide deeper insight into the central role of system prompts in bias propagation and offer a practical, deployable approach for building more socially responsible T2I systems.</p> + https://arxiv.org/abs/2512.04981 + Thu, 04 Dec 2025 16:52:45 +0000 - MIRA: Multimodal Iterative Reasoning Agent for Image Editing - https://arxiv.org/abs/2511.21087 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.21087.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Ziyun Zeng, Hang Hua, Jiebo Luo</p><p><b>Upvotes:</b> 9</p><p><b>Summary:</b> Instruction-guided image editing offers an intuitive way for users to edit images with natural language. However, diffusion-based editing models often struggle to accurately interpret complex user instructions, especially those involving compositional relationships, contextual cues, or referring expressions, leading to edits that drift semantically or fail to reflect the intended changes. We tackle this problem by proposing MIRA (Multimodal Iterative Reasoning Agent), a lightweight, plug-and-play multimodal reasoning agent that performs editing through an iterative perception-reasoning-action loop, effectively simulating multi-turn human-model interaction processes. Instead of issuing a single prompt or static plan, MIRA predicts atomic edit instructions step by step, using visual feedback to make its decisions. Our 150K multimodal tool-use dataset, MIRA-Editing, combined with a two-stage SFT + GRPO training pipeline, enables MIRA to perform reasoning and editing over complex editing instructions. When paired with open-source image editing models such as Flux.1-Kontext, Step1X-Edit, and Qwen-Image-Edit, MIRA significantly improves both semantic consistency and perceptual quality, achieving performance comparable to or exceeding proprietary systems such as GPT-Image and Nano-Banana.</p> - https://arxiv.org/abs/2511.21087 - Wed, 26 Nov 2025 06:13:32 +0000 + Nex-N1: Agentic Models Trained via a Unified Ecosystem for Large-Scale Environment Construction + https://arxiv.org/abs/2512.04987 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2512.04987.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Nex-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</p><p><b>Upvotes:</b> 62</p><p><b>Summary:</b> 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.</p> + https://arxiv.org/abs/2512.04987 + Thu, 04 Dec 2025 16:57:02 +0000 - Monet: Reasoning in Latent Visual Space Beyond Images and Language - https://arxiv.org/abs/2511.21395 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.21395.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Qixun Wang, Yang Shi, Yifei Wang, Yuanxing Zhang, Pengfei Wan, Kun Gai, Xianghua Ying, Yisen Wang</p><p><b>Upvotes:</b> 13</p><p><b>Summary:</b> "Thinking with images" has emerged as an effective paradigm for advancing visual reasoning, extending beyond text-only chains of thought by injecting visual evidence into intermediate reasoning steps. However, existing methods fall short of human-like abstract visual thinking, as their flexibility is fundamentally limited by external tools. In this work, we introduce Monet, a training framework that enables multimodal large language models (MLLMs) to reason directly within the latent visual space by generating continuous embeddings that function as intermediate visual thoughts. We identify two core challenges in training MLLMs for latent visual reasoning: high computational cost in latent-vision alignment and insufficient supervision over latent embeddings, and address them with a three-stage distillation-based supervised fine-tuning (SFT) pipeline. We further reveal a limitation of applying GRPO to latent reasoning: it primarily enhances text-based reasoning rather than latent reasoning. To overcome this, we propose VLPO (Visual-latent Policy Optimization), a reinforcement learning method that explicitly incorporates latent embeddings into policy gradient updates. To support SFT, we construct Monet-SFT-125K, a high-quality text-image interleaved CoT dataset containing 125K real-world, chart, OCR, and geometry CoTs. Our model, Monet-7B, shows consistent gains across real-world perception and reasoning benchmarks and exhibits strong out-of-distribution generalization on challenging abstract visual reasoning tasks. We also empirically analyze the role of each training component and discuss our early unsuccessful attempts, providing insights for future developments in visual latent reasoning. Our model, data, and code are available at https://github.com/NOVAglow646/Monet.</p> - https://arxiv.org/abs/2511.21395 - Wed, 26 Nov 2025 13:46:39 +0000 + Reflection Removal through Efficient Adaptation of Diffusion Transformers + https://arxiv.org/abs/2512.05000 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2512.05000.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Daniyar Zakarin, Thiemo Wandel, Anton Obukhov, Dengxin Dai</p><p><b>Upvotes:</b> 11</p><p><b>Summary:</b> We introduce a diffusion-transformer (DiT) framework for single-image reflection removal that leverages the generalization strengths of foundation diffusion models in the restoration setting. Rather than relying on task-specific architectures, we repurpose a pre-trained DiT-based foundation model by conditioning it on reflection-contaminated inputs and guiding it toward clean transmission layers. We systematically analyze existing reflection removal data sources for diversity, scalability, and photorealism. To address the shortage of suitable data, we construct a physically based rendering (PBR) pipeline in Blender, built around the Principled BSDF, to synthesize realistic glass materials and reflection effects. Efficient LoRA-based adaptation of the foundation model, combined with the proposed synthetic data, achieves state-of-the-art performance on in-domain and zero-shot benchmarks. These results demonstrate that pretrained diffusion transformers, when paired with physically grounded data synthesis and efficient adaptation, offer a scalable and high-fidelity solution for reflection removal. Project page: https://hf.co/spaces/huawei-bayerlab/windowseat-reflection-removal-web</p> + https://arxiv.org/abs/2512.05000 + Thu, 04 Dec 2025 17:12:39 +0000 - Video Generation Models Are Good Latent Reward Models - https://arxiv.org/abs/2511.21541 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.21541.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Xiaoyue Mi, Wenqing Yu, Jiesong Lian, Shibo Jie, Ruizhe Zhong, Zijun Liu, Guozhen Zhang, Zixiang Zhou, Zhiyong Xu, Yuan Zhou, Qinglin Lu, Fan Tang</p><p><b>Upvotes:</b> 32</p><p><b>Summary:</b> Reward feedback learning (ReFL) has proven effective for aligning image generation with human preferences. However, its extension to video generation faces significant challenges. Existing video reward models rely on vision-language models designed for pixel-space inputs, confining ReFL optimization to near-complete denoising steps after computationally expensive VAE decoding. This pixel-space approach incurs substantial memory overhead and increased training time, and its late-stage optimization lacks early-stage supervision, refining only visual quality rather than fundamental motion dynamics and structural coherence. In this work, we show that pre-trained video generation models are naturally suited for reward modeling in the noisy latent space, as they are explicitly designed to process noisy latent representations at arbitrary timesteps and inherently preserve temporal information through their sequential modeling capabilities. Accordingly, we propose Process Reward Feedback Learning~(PRFL), a framework that conducts preference optimization entirely in latent space, enabling efficient gradient backpropagation throughout the full denoising chain without VAE decoding. Extensive experiments demonstrate that PRFL significantly improves alignment with human preferences, while achieving substantial reductions in memory consumption and training time compared to RGB ReFL.</p> - https://arxiv.org/abs/2511.21541 - Wed, 26 Nov 2025 16:14:18 +0000 + 4DLangVGGT: 4D Language-Visual Geometry Grounded Transformer + https://arxiv.org/abs/2512.05060 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2512.05060.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Xianfeng Wu, Yajing Bai, Minghan Li, Xianzu Wu, Xueqi Zhao, Zhongyuan Lai, Wenyu Liu, Xinggang Wang</p><p><b>Upvotes:</b> 17</p><p><b>Summary:</b> 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</p> + https://arxiv.org/abs/2512.05060 + Thu, 04 Dec 2025 18:15:27 +0000 - Harmony: Harmonizing Audio and Video Generation through Cross-Task Synergy - https://arxiv.org/abs/2511.21579 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.21579.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Teng Hu, Zhentao Yu, Guozhen Zhang, Zihan Su, Zhengguang Zhou, Youliang Zhang, Yuan Zhou, Qinglin Lu, Ran Yi</p><p><b>Upvotes:</b> 20</p><p><b>Summary:</b> The synthesis of synchronized audio-visual content is a key challenge in generative AI, with open-source models facing challenges in robust audio-video alignment. Our analysis reveals that this issue is rooted in three fundamental challenges of the joint diffusion process: (1) Correspondence Drift, where concurrently evolving noisy latents impede stable learning of alignment; (2) inefficient global attention mechanisms that fail to capture fine-grained temporal cues; and (3) the intra-modal bias of conventional Classifier-Free Guidance (CFG), which enhances conditionality but not cross-modal synchronization. To overcome these challenges, we introduce Harmony, a novel framework that mechanistically enforces audio-visual synchronization. We first propose a Cross-Task Synergy training paradigm to mitigate drift by leveraging strong supervisory signals from audio-driven video and video-driven audio generation tasks. Then, we design a Global-Local Decoupled Interaction Module for efficient and precise temporal-style alignment. Finally, we present a novel Synchronization-Enhanced CFG (SyncCFG) that explicitly isolates and amplifies the alignment signal during inference. Extensive experiments demonstrate that Harmony establishes a new state-of-the-art, significantly outperforming existing methods in both generation fidelity and, critically, in achieving fine-grained audio-visual synchronization.</p> - https://arxiv.org/abs/2511.21579 - Wed, 26 Nov 2025 16:53:05 +0000 + Deep Forcing: Training-Free Long Video Generation with Deep Sink and Participative Compression + https://arxiv.org/abs/2512.05081 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2512.05081.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Jung Yi, Wooseok Jang, Paul Hyunbin Cho, Jisu Nam, Heeji Yoon, Seungryong Kim</p><p><b>Upvotes:</b> 19</p><p><b>Summary:</b> 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.</p> + https://arxiv.org/abs/2512.05081 + Thu, 04 Dec 2025 18:46:44 +0000 - Multi-Crit: Benchmarking Multimodal Judges on Pluralistic Criteria-Following - https://arxiv.org/abs/2511.21662 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.21662.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Tianyi Xiong, Yi Ge, Ming Li, Zuolong Zhang, Pranav Kulkarni, Kaishen Wang, Qi He, Zeying Zhu, Chenxi Liu, Ruibo Chen, Tong Zheng, Yanshuo Chen, Xiyao Wang, Renrui Zhang, Wenhu Chen, Heng Huang</p><p><b>Upvotes:</b> 9</p><p><b>Summary:</b> Large multimodal models (LMMs) are increasingly adopted as judges in multimodal evaluation systems due to their strong instruction following and consistency with human preferences. However, their ability to follow diverse, fine-grained evaluation criteria remains underexplored. We develop Multi-Crit, a benchmark for evaluating multimodal judges on their capacity to follow pluralistic criteria and produce reliable criterion-level judgments. Covering both open-ended generation and verifiable reasoning tasks, Multi-Crit is built through a rigorous data curation pipeline that gathers challenging response pairs with multi-criterion human annotations. It further introduces three novel metrics for systematically assessing pluralistic adherence, criterion-switching flexibility, and the ability to recognize criterion-level preference conflicts. Comprehensive analysis of 25 LMMs reveals that 1) proprietary models still struggle to maintain consistent adherence to pluralistic criteria--especially in open-ended evaluation; 2) open-source models lag further behind in flexibly following diverse criteria; and 3) critic fine-tuning with holistic judgment signals enhances visual grounding but fails to generalize to pluralistic criterion-level judgment. Additional analyses on reasoning fine-tuning, test-time scaling, and boundary consistency between open-source and proprietary models further probe the limits of current multimodal judges. As a pioneering study, Multi-Crit lays the foundation for building reliable and steerable multimodal AI evaluation.</p> - https://arxiv.org/abs/2511.21662 - Wed, 26 Nov 2025 18:35:17 +0000 + TV2TV: A Unified Framework for Interleaved Language and Video Generation + https://arxiv.org/abs/2512.05103 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2512.05103.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Xiaochuang Han, Youssef Emad, Melissa Hall, John Nguyen, Karthik Padthe, Liam Robbins, Amir Bar, Delong Chen, Michal Drozdzal, Maha Elbayad, Yushi Hu, Shang-Wen Li, Sreya Dutta Roy, Jakob Verbeek, XuDong Wang, Marjan Ghazvininejad, Luke Zettlemoyer, Emily Dinan</p><p><b>Upvotes:</b> 10</p><p><b>Summary:</b> Video generation models are rapidly advancing, but can still struggle with complex video outputs that require significant semantic branching or repeated high-level reasoning about what should happen next. In this paper, we introduce a new class of omni video-text models that integrate ideas from recent LM reasoning advances to address this challenge. More specifically, we present TV2TV, a unified generative modeling framework which decomposes video generation into an interleaved text and video generation process. TV2TV jointly learns language modeling (next-token prediction) and video flow matching (next-frame prediction) using a Mixture-of-Transformers (MoT) architecture. At inference time, TV2TV decides when to alternate between generating text and video frames, allowing the model to "think in words" about subsequent content before ``acting in pixels'' to produce frames. This design offloads much of the responsibility for deciding what should happen next to the language modeling tower, enabling improved visual quality and prompt alignment of generated videos. It also enables fine-grained controllability, allowing users to modify the video generation trajectory through text interventions at any point in the process. In controlled experiments on video game data, TV2TV demonstrates substantial improvements in both visual quality and controllability. TV2TV also scales to natural videos, as we show by augmenting sports videos with interleaved natural language action descriptions using vision-language models (VLMs). Training TV2TV on this corpus yields strong visual quality and prompt alignment, showcasing the model's ability to reason about and generate complex real-world action sequences. Together, these results highlight TV2TV as a promising step toward video generation with open-ended textual reasoning and control.</p> + https://arxiv.org/abs/2512.05103 + Thu, 04 Dec 2025 18:59:09 +0000 - Agentic Learner with Grow-and-Refine Multimodal Semantic Memory - https://arxiv.org/abs/2511.21678 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.21678.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Weihao Bo, Shan Zhang, Yanpeng Sun, Jingjing Wu, Qunyi Xie, Xiao Tan, Kunbin Chen, Wei He, Xiaofan Li, Na Zhao, Jingdong Wang, Zechao Li</p><p><b>Upvotes:</b> 8</p><p><b>Summary:</b> MLLMs exhibit strong reasoning on isolated queries, yet they operate de novo -- solving each problem independently and often repeating the same mistakes. Existing memory-augmented agents mainly store past trajectories for reuse. However, trajectory-based memory suffers from brevity bias, gradually losing essential domain knowledge. More critically, even in truly multimodal problem-solving settings, it records only a single-modality trace of past behavior, failing to preserve how visual attention and logical reasoning jointly contributed to the solution. This is fundamentally misaligned with human cognition: semantic memory is both multimodal and integrated, preserving visual and abstract knowledge through coordinated but distinct representational streams. We thus introduce ViLoMem, a dual-stream memory framework that constructs compact, schema-based memory. It separately encodes visual distraction patterns and logical reasoning errors, enabling MLLMs to learn from their successful and failed experiences. Following a grow-and-refine principle, the system incrementally accumulates and updates multimodal semantic knowledge -- preserving stable, generalizable strategies while avoiding catastrophic forgetting. Across six multimodal benchmarks, ViLoMem consistently improves pass@1 accuracy and substantially reduces repeated visual and logical errors. Ablations confirm the necessity of dual-stream memory with explicit distraction--hallucination separation, demonstrating the value of error-aware multimodal memory for lifelong and cross-domain agentic learning. Our project page will be available at https://weihao-bo.github.io/ViLoMeo-page.</p> - https://arxiv.org/abs/2511.21678 - Wed, 26 Nov 2025 18:55:08 +0000 + NeuralRemaster: Phase-Preserving Diffusion for Structure-Aligned Generation + https://arxiv.org/abs/2512.05106 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2512.05106.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Yu Zeng, Charles Ochoa, Mingyuan Zhou, Vishal M. Patel, Vitor Guizilini, Rowan McAllister</p><p><b>Upvotes:</b> 11</p><p><b>Summary:</b> 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 https://yuzeng-at-tri.github.io/ppd-page/{project page}.</p> + https://arxiv.org/abs/2512.05106 + Thu, 04 Dec 2025 18:59:18 +0000 - G^2VLM: Geometry Grounded Vision Language Model with Unified 3D Reconstruction and Spatial Reasoning - https://arxiv.org/abs/2511.21688 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.21688.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Wenbo Hu, Jingli Lin, Yilin Long, Yunlong Ran, Lihan Jiang, Yifan Wang, Chenming Zhu, Runsen Xu, Tai Wang, Jiangmiao Pang</p><p><b>Upvotes:</b> 8</p><p><b>Summary:</b> Vision-Language Models (VLMs) still lack robustness in spatial intelligence, demonstrating poor performance on spatial understanding and reasoning tasks. We attribute this gap to the absence of a visual geometry learning process capable of reconstructing 3D space from 2D images. We present G^2VLM, a geometry grounded vision-language model that bridges two fundamental aspects of spatial intelligence: spatial 3D reconstruction and spatial understanding. G^2VLM natively leverages learned 3D visual geometry features to directly predict 3D attributes and enhance spatial reasoning tasks via in-context learning and interleaved reasoning. Our unified design is highly scalable for spatial understanding: it trains on abundant multi-view image and video data, while simultaneously leveraging the benefits of 3D visual priors that are typically only derived from hard-to-collect annotations. Experimental results demonstrate G^2VLM is proficient in both tasks, achieving comparable results to state-of-the-art feed-forward 3D reconstruction models and achieving better or competitive results across spatial understanding and reasoning tasks. By unifying a semantically strong VLM with low-level 3D vision tasks, we hope G^2VLM can serve as a strong baseline for the community and unlock more future applications, such as 3D scene editing.</p> - https://arxiv.org/abs/2511.21688 - Wed, 26 Nov 2025 18:59:39 +0000 + ARM-Thinker: Reinforcing Multimodal Generative Reward Models with Agentic Tool Use and Visual Reasoning + https://arxiv.org/abs/2512.05111 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2512.05111.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Shengyuan Ding, Xinyu Fang, Ziyu Liu, Yuhang Zang, Yuhang Cao, Xiangyu Zhao, Haodong Duan, Xiaoyi Dong, Jianze Liang, Bin Wang, Conghui He, Dahua Lin, Jiaqi Wang</p><p><b>Upvotes:</b> 42</p><p><b>Summary:</b> 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.</p> + https://arxiv.org/abs/2512.05111 + Thu, 04 Dec 2025 18:59:52 +0000 - Canvas-to-Image: Compositional Image Generation with Multimodal Controls - https://arxiv.org/abs/2511.21691 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.21691.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Yusuf Dalva, Guocheng Gordon Qian, Maya Goldenberg, Tsai-Shien Chen, Kfir Aberman, Sergey Tulyakov, Pinar Yanardag, Kuan-Chieh Jackson Wang</p><p><b>Upvotes:</b> 22</p><p><b>Summary:</b> While modern diffusion models excel at generating high-quality and diverse images, they still struggle with high-fidelity compositional and multimodal control, particularly when users simultaneously specify text prompts, subject references, spatial arrangements, pose constraints, and layout annotations. We introduce Canvas-to-Image, a unified framework that consolidates these heterogeneous controls into a single canvas interface, enabling users to generate images that faithfully reflect their intent. Our key idea is to encode diverse control signals into a single composite canvas image that the model can directly interpret for integrated visual-spatial reasoning. We further curate a suite of multi-task datasets and propose a Multi-Task Canvas Training strategy that optimizes the diffusion model to jointly understand and integrate heterogeneous controls into text-to-image generation within a unified learning paradigm. This joint training enables Canvas-to-Image to reason across multiple control modalities rather than relying on task-specific heuristics, and it generalizes well to multi-control scenarios during inference. Extensive experiments show that Canvas-to-Image significantly outperforms state-of-the-art methods in identity preservation and control adherence across challenging benchmarks, including multi-person composition, pose-controlled composition, layout-constrained generation, and multi-control generation.</p> - https://arxiv.org/abs/2511.21691 - Wed, 26 Nov 2025 18:59:56 +0000 + DraCo: Draft as CoT for Text-to-Image Preview and Rare Concept Generation + https://arxiv.org/abs/2512.05112 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2512.05112.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Dongzhi Jiang, Renrui Zhang, Haodong Li, Zhuofan Zong, Ziyu Guo, Jun He, Claire Guo, Junyan Ye, Rongyao Fang, Weijia Li, Rui Liu, Hongsheng Li</p><p><b>Upvotes:</b> 9</p><p><b>Summary:</b> 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.</p> + https://arxiv.org/abs/2512.05112 + Thu, 04 Dec 2025 18:59:53 +0000 - Revisiting Generalization Across Difficulty Levels: It's Not So Easy - https://arxiv.org/abs/2511.21692 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2511.21692.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Yeganeh Kordi, Nihal V. Nayak, Max Zuo, Ilana Nguyen, Stephen H. Bach</p><p><b>Upvotes:</b> 13</p><p><b>Summary:</b> We investigate how well large language models (LLMs) generalize across different task difficulties, a key question for effective data curation and evaluation. Existing research is mixed regarding whether training on easier or harder data leads to better results, and whether those gains come on easier or harder test data. We address this question by conducting a systematic evaluation of LLMs' generalization across models, datasets, and fine-grained groups of example difficulty. We rank examples in six datasets using the outputs of thousands of different LLMs and Item Response Theory (IRT), a well-established difficulty metric in educational testing. Unlike prior work, our difficulty ratings are therefore determined solely by the abilities of many different LLMs, excluding human opinions of difficulty. With a more objective, larger-scale, and finer-grained analysis, we show that cross-difficulty generalization is often limited; training on either easy or hard data cannot achieve consistent improvements across the full range of difficulties. These results show the importance of having a range of difficulties in both training and evaluation data for LLMs, and that taking shortcuts with respect to difficulty is risky.</p> - https://arxiv.org/abs/2511.21692 - Wed, 26 Nov 2025 18:59:57 +0000 + Splannequin: Freezing Monocular Mannequin-Challenge Footage with Dual-Detection Splatting + https://arxiv.org/abs/2512.05113 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2512.05113.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Hao-Jen Chien, Yi-Chuan Huang, Chung-Ho Wu, Wei-Lun Chao, Yu-Lun Liu</p><p><b>Upvotes:</b> 14</p><p><b>Summary:</b> 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/</p> + https://arxiv.org/abs/2512.05113 + Thu, 04 Dec 2025 18:59:53 +0000