{"ok":true,"snapshot":{"date":"2026-08-02","capturedAt":"2026-08-02T14:15:33.394Z","total_papers":30,"raw_count":50,"papers":[{"paperId":"2607.28618","title":"AskChem: Claim-Centered Infrastructure for Chemistry Literature Synthesis","summary":"Chemistry literature synthesis often requires assembling specific findings scattered across many publications, yet existing literature-search systems primarily return ranked document lists. As a result, scientists and AI agents need to locate relevant information, verify their provenance, and assemble cross-paper answers manually. We present AskChem, a claim-centered infrastructure for cross-paper chemistry search. AskChem changes the unit of retrieval from the paper to the provenance-carrying claim: each paper is converted into atomic, typed claims, each grounded by a source DOI and a verbatim quote or an explicit evidence locator. Over this shared claim store, AskChem exposes complementary structures for search and synthesis: a stabilized faceted taxonomy for hierarchical retrieval and …","authors":["Bing Yan","Gregory Wolfe","Stefano Martiniani","Kyunghyun Cho"],"publishedAt":"2026-07-30T00:00:00.000Z","submittedAt":null,"upvotes":292,"num_comments":2,"thumbnail":"https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2607.28618.png","hf_url":"https://huggingface.co/papers/2607.28618","arxiv_url":"https://arxiv.org/abs/2607.28618","github_repo":null,"github_stars":null,"ai_keywords":[]},{"paperId":"2607.28227","title":"Qwen-UI-Agent Technical Report: Toward Next-Generation Real-World Centric Foundation GUI Agents","summary":"GUI agents have the potential to become a general purpose executor over existing digital devices. To advance them toward real-world use, we envision agents that operate reliably on real devices, execute workflows across platforms, combine GUI interaction with CLI execution, complete long-horizon tasks, proactively initiate useful services, and autonomously improve their capabilities with minimal human effort. Guided by this vision, we present Qwen-UI-Agent, a real-world centric foundation GUI agent spanning mobile, computer-use, web, and DeepSearch environments. Qwen-UI-Agent combines diverse sandbox environments with a large-scale real-device mobile runtime. Its unified action space interleaves GUI operations with CLI execution and generates batched actions in a single model turn. An Aut…","authors":["Hanzhang Zhou","Panrong Tong","Xu Zhang","Quyu Kong","Chenglin Cai","Tianyu Xia","Gongjie Zhang","Jianan Zhang"],"publishedAt":"2026-07-30T00:00:00.000Z","submittedAt":null,"upvotes":279,"num_comments":7,"thumbnail":"https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2607.28227.png","hf_url":"https://huggingface.co/papers/2607.28227","arxiv_url":"https://arxiv.org/abs/2607.28227","github_repo":null,"github_stars":null,"ai_keywords":[]},{"paperId":"2607.26760","title":"Metis: Memory Foundation Model","summary":"Recent advances in AI agents have increasingly internalized native capabilities into their underlying foundation models, giving rise to multimodal foundation models and large reasoning models. However, agent memory is still primarily implemented through external modules, leaving the native memory capability largely unexplored. In this paper, we take a first step toward this direction by introducing memory foundation models, which empower foundation models with native memory capabilities. We formalize native memory from two perspectives: a persistent and dynamically evolving memory state within the backbone, and native memory procedures that autonomously store and utilize information through model computation. We show that native memory offers advantages in architecture, end-to-end optimiz…","authors":["Zeyu Zhang","Ziliang Guo","Yihang Sun","Xichong Zhang","Xixuan Hao","Zehao Lin","Yang Zhang","Xiaoyan Zhao"],"publishedAt":"2026-07-29T00:00:00.000Z","submittedAt":null,"upvotes":256,"num_comments":3,"thumbnail":"https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2607.26760.png","hf_url":"https://huggingface.co/papers/2607.26760","arxiv_url":"https://arxiv.org/abs/2607.26760","github_repo":null,"github_stars":null,"ai_keywords":[]},{"paperId":"2607.28568","title":"Frontis-MA1: Training an AI4AI Model towards Recursive Self-Improvement in Machine Learning Engineering","summary":"Recursive self-improvement (RSI) requires AI systems that improve the process of building AI (i.e., AI4AI); machine learning engineering (MLE) offers a concrete, executable testbed for studying this capability. We introduce OpenMLE, an open full-stack system for RSI research in MLE, spanning verifiable task environments with execution feedback (OpenMLE-Gym), operator learning (OpenMLE-RL), and long-horizon search (OpenMLE-Evo). On this stack we post-train Frontis-MA1 (35B) as a meta-evolution agent for MLE, aligning post-training and inference around four atomic program-evolution operators (Draft, Improve, Debug, Crossover): the same operators are trained via execution-grounded SFT and RL on data deduplicated against all evaluation benchmarks, then composed into long-horizon search, coupl…","authors":["Junlin Yang","Che Jiang","Yu Fu","Tianwei Luo","Can Ren","Weizhi Wang","Kaikai Zhao","Hongyi Liu"],"publishedAt":"2026-07-30T00:00:00.000Z","submittedAt":null,"upvotes":166,"num_comments":2,"thumbnail":"https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2607.28568.png","hf_url":"https://huggingface.co/papers/2607.28568","arxiv_url":"https://arxiv.org/abs/2607.28568","github_repo":null,"github_stars":null,"ai_keywords":[]},{"paperId":"2607.28624","title":"PhiZero: A World Model Built Around Physical Language","summary":"We introduce PhiZero, a physical world model built around physical language, a compact discrete representation of world-state transitions. Existing physical world models typically predict future videos directly in pixel space, leaving the underlying world dynamics implicit within high-dimensional visual predictors. Motivated by humans' ability to abstract predictive structure from visual experience and organize it in natural language for explicit reasoning, we learn physical language from in-the-wild videos through self-supervision and use it to explicitly reason about how the physical world evolves. Accordingly, PhiZero adopts a reason-then-render paradigm: it first infers future world evolution as a physical-language sequence and then renders the inferred transitions into videos. Extens…","authors":["Shuyao Shang","Yuqi Wang","Ruopeng Gao","Xu Chen","Tieniu Tan","Lue Fan","Zhaoxiang Zhang"],"publishedAt":"2026-07-30T00:00:00.000Z","submittedAt":null,"upvotes":157,"num_comments":3,"thumbnail":"https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2607.28624.png","hf_url":"https://huggingface.co/papers/2607.28624","arxiv_url":"https://arxiv.org/abs/2607.28624","github_repo":null,"github_stars":null,"ai_keywords":[]},{"paperId":"2607.26811","title":"DistillAlign: Coordinating Mode Covering and Mode Seeking in Autoregressive Video Distillation","summary":"Existing autoregressive video distillation methods commonly adopt a Distribution Matching Distillation (DMD)-based multi-stage pipeline. However, they typically decouple the initialization and DMD stages -- which then pursue different target distributions -- and judge the intermediate student mainly by visual scores such as VBench. In this paper, we revisit this design from a distributional perspective. Given the mode-seeking nature of the distribution matching loss, a good initialization should match the mode coverage of the target DMD teacher, rather than merely pursuing high quality. To analyze this, we introduce a distributional evaluation protocol that measures precision and coverage between student and teacher distributions in a shared latent space. It exposes differences hidden by …","authors":["Jiaxing Li","Kai Zou","Cindy Zhou","Kaichen Huang","Junyao Gao","Zile Wang","Yang Liu","Bin Liu"],"publishedAt":"2026-07-29T00:00:00.000Z","submittedAt":null,"upvotes":89,"num_comments":2,"thumbnail":"https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2607.26811.png","hf_url":"https://huggingface.co/papers/2607.26811","arxiv_url":"https://arxiv.org/abs/2607.26811","github_repo":null,"github_stars":null,"ai_keywords":[]},{"paperId":"2607.27380","title":"VideoCoCo: Code-as-CoT for Physically-Consistent Video Generation via an Agentic Dual-Engine System","summary":"Text-to-video models have achieved remarkable visual quality, yet they still struggle to generate physically consistent dynamics because the temporal evolution of a scene must be inferred implicitly from a highly compressed text prompt. Existing chain-of-thought approaches introduce intermediate plans or visual states, but these representations are typically non-executable or temporally sparse, limiting their ability to instantiate and control the complete spatiotemporal process. To address this limitation, we introduce VideoCoCo, an agentic dual-engine framework in which executable Blender code serves as a process-level chain of thought. Given a text prompt, a coding agent synthesizes a Blender program that explicitly specifies the scene and its temporal evolution. The executable simulat…","authors":["Haodong Li","Tianfei Ren","Xiaoxiao Ma","Chunmei Qing","Zhen Fang","Sipeng He","Ziyu Guo","Haoyu Wu"],"publishedAt":"2026-07-29T00:00:00.000Z","submittedAt":null,"upvotes":64,"num_comments":2,"thumbnail":"https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2607.27380.png","hf_url":"https://huggingface.co/papers/2607.27380","arxiv_url":"https://arxiv.org/abs/2607.27380","github_repo":null,"github_stars":null,"ai_keywords":[]},{"paperId":"2607.27919","title":"Memory Decoder at Scale: A Pretrained, Parametric Long-Term Memory","summary":"Decoder-only language models entangle long-term memory and reasoning in a single parameter set, making it difficult to scale memory capacity independently. Memory Decoder introduces a parametric long-term memory module but only studies it at a relatively small scale. In this work, we present Memory Decoder at Scale, scaling memory models up to 6.9B parameters and pretraining them on 300B tokens. At this data scale, the combined cost of indexing and search makes a standard Faiss pipeline infeasible. We address this bottleneck with a distributed pipeline for Faiss indexing and retrieval, together with sparse, batch-wise loading of kNN distributions. Across model scales, we find that allocating more parameters to memory yields a better parameter-performance tradeoff than scaling the base mod…","authors":["Rubin Wei","Jiaqi Cao","Jiarui Wang","Junming Zhang","Qipeng Guo","Bowen Zhou","Zhouhan Lin"],"publishedAt":"2026-07-30T00:00:00.000Z","submittedAt":null,"upvotes":49,"num_comments":2,"thumbnail":"https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2607.27919.png","hf_url":"https://huggingface.co/papers/2607.27919","arxiv_url":"https://arxiv.org/abs/2607.27919","github_repo":null,"github_stars":null,"ai_keywords":[]},{"paperId":"2607.28595","title":"Beacon: Knowing When and How to Perform Agentic Visual Reasoning","summary":"The fundamental goal of agentic visual reasoning is to improve the success rate of multimodal large language models (MLLMs) on complex tasks, rather than merely equipping them with a sophisticated yet inefficient reasoning paradigm. In this work, we rethink agentic visual reasoning through two key dimensions of tool use: Mode Adaptiveness (MA) and Tool Effect (TE). Mode Adaptiveness characterizes whether an MLLM can recognize when tools are truly necessary and invoke them accordingly, thereby avoiding unnecessary computational overhead while improving performance on challenging problems that require tool assistance. Tool Effect characterizes the actual impact of tool use: tools should extend the model's capabilities on problems unsolvable through text-only reasoning, while avoiding additi…","authors":["Qixun Wang","Yang Shi","Letian Cheng","Zhuoran Zhang","Yan He","Yuqi Tang","Qi Zhang","Xinlei Yu"],"publishedAt":"2026-07-30T00:00:00.000Z","submittedAt":null,"upvotes":46,"num_comments":2,"thumbnail":"https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2607.28595.png","hf_url":"https://huggingface.co/papers/2607.28595","arxiv_url":"https://arxiv.org/abs/2607.28595","github_repo":null,"github_stars":null,"ai_keywords":[]},{"paperId":"2607.26497","title":"BM25 Wins at Scale: A Scaling Study of Retrieval-Augmented Generation Paradigms","summary":"Retrieval-augmented generation (RAG) spans lexical and dense retrieval, graph-based indexing, and agentic search, but these paradigms are usually evaluated on different benchmarks at one corpus size, leaving their accuracy-cost scaling unclear. To bridge this gap, we present a controlled study that varies corpus size along 28 strictly nested tiers spanning roughly 450-fold, while holding questions and a fixed bedrock of relevant and adversarial documents unchanged. Under one reader model and one judging protocol, we measure official accuracy, construction and query tokens, and latency. The results reveal a scale-dependent crossover rather than an unconditional winner. File-System Agent leads at the smallest shared tiers, but its sequential exploration costs 39 times more query tokens at t…","authors":["Pengyu Wang","Benfeng Xu","Shaohan Wang","Xin Zeng","Huarui Wu","Lei Zhang","Licheng Zhang"],"publishedAt":"2026-07-30T00:00:00.000Z","submittedAt":null,"upvotes":44,"num_comments":4,"thumbnail":"https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2607.26497.png","hf_url":"https://huggingface.co/papers/2607.26497","arxiv_url":"https://arxiv.org/abs/2607.26497","github_repo":null,"github_stars":null,"ai_keywords":[]},{"paperId":"2607.25294","title":"CLBench-V: Evaluating Multimodal Context Learning from Grounding to Knowledge Acquisition","summary":"Real-world tasks often require models to learn from task-specific context rather than relying only on pre-trained knowledge. While recent work has highlighted this capability as context learning, existing evaluations mainly focus on textual contexts. In many practical settings, however, the context to be learned from is multimodal: scientific findings are conveyed through figures and tables, financial indicators are scattered across converted reports, and spatial decisions depend on maps, scenes, or web pages. We introduce CLBench-V, a benchmark for multimodal context learning that addresses the difficulty of localizing where context use breaks down by organizing tasks around three dimensions: context grounding, new information application, and new knowledge learning. CLBench-V combines c…","authors":["Lai Wei","Chengqi Li","Jiapeng Li","Ruina Hu","Yue Wang","Weiran Huang"],"publishedAt":"2026-07-28T00:00:00.000Z","submittedAt":null,"upvotes":44,"num_comments":2,"thumbnail":"https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2607.25294.png","hf_url":"https://huggingface.co/papers/2607.25294","arxiv_url":"https://arxiv.org/abs/2607.25294","github_repo":null,"github_stars":null,"ai_keywords":[]},{"paperId":"2607.28022","title":"Flux-OPD: On-Policy Distillation with Evolving Contexts","summary":"Large language model training in open-ended domains lacks verifiable rewards, making task preferences difficult to formalize as effective supervision. Contexts can convey such preferences, yet provide little additional supervision once distilled into the student, motivating contexts that evolve with student performance. However, directly using evolving contexts as in-training supervision results in an unstable distillation target and conflicting distributions, requiring mechanisms to stabilize target and downweight conflicts. In this paper, we analyze the effect of contexts through a decomposition of the reverse KL objective, revealing two findings: the student is distilled toward the geometric mean of context-conditioned teachers, and the objective contains a conflict term that measures …","authors":["Yuran Wang","Zekun Wang","Bohan Zeng","Ruixu Zhang","Wenxuan Liu","Liu Yang","Yifan Dai","Yang Shi"],"publishedAt":"2026-07-30T00:00:00.000Z","submittedAt":null,"upvotes":39,"num_comments":2,"thumbnail":"https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2607.28022.png","hf_url":"https://huggingface.co/papers/2607.28022","arxiv_url":"https://arxiv.org/abs/2607.28022","github_repo":null,"github_stars":null,"ai_keywords":[]},{"paperId":"2607.27616","title":"MPIE-Bench: Benchmarking Anatomically Plausible Multi-Person Interaction Editing","summary":"Text-to-image and personalized editing models now synthesize high-fidelity single-subject images with ease. Yet placing multiple named people into shared contact actions such as embrace, carry, or grapple still exposes major failures: fused limbs, invented extremities, and interpenetrating bodies. Existing evaluations largely overlook these anatomical and geometric issues, and VLM-as-a-judge checklists often saturate on Interaction while the errors remain obvious to humans. We introduce MPIE-Bench, a 2,500-sample benchmark of video-mined editing triplets spanning 405 scenes, 14 interaction categories, and four contact densities (C0-C3). We also propose MPIE-Eval, whose two new axes score contact-time geometry from a frozen public multi-person mesh reconstruction. Anatomy asks whether ever…","authors":["Jiajia Lin","Mingxuan Du","Tuowen Zhou","Benfeng Xu","Hongtao Xie"],"publishedAt":"2026-07-30T00:00:00.000Z","submittedAt":null,"upvotes":37,"num_comments":2,"thumbnail":"https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2607.27616.png","hf_url":"https://huggingface.co/papers/2607.27616","arxiv_url":"https://arxiv.org/abs/2607.27616","github_repo":null,"github_stars":null,"ai_keywords":[]},{"paperId":"2607.28625","title":"ACE-Data-0: Human-Centric Ambient Capture as Embodied Data Engine","summary":"Embodied intelligence faces a fundamental data bottleneck. Models must capture how first-person perception, whole-body motion, dexterous manipulation, object state, sound, and touch evolve together as humans pursue goals over time. Existing datasets fragment this experience across viewpoints, modalities, or spatial scales, leaving the full perception-action loop only partially observed. We introduce the Ambient Capture Engine (ACE), a human-centric data engine that transforms real home environments into spatially calibrated, temporally synchronized recording studios. ACE operates at two complementary scales: a table-scale configuration resolves hand-object manipulation, while a room-scale configuration captures whole-body motion, locomotion, and interactions across a furnished home. ACE r…","authors":["Yukang Cao","Haozhe Xie","Beichen Wen","Runmao Yao","Yinghao Liu","Yue Huang","Zhichao Liao","Yunxiang Wang"],"publishedAt":"2026-07-30T00:00:00.000Z","submittedAt":null,"upvotes":35,"num_comments":1,"thumbnail":"https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2607.28625.png","hf_url":"https://huggingface.co/papers/2607.28625","arxiv_url":"https://arxiv.org/abs/2607.28625","github_repo":null,"github_stars":null,"ai_keywords":[]},{"paperId":"2607.27816","title":"Beyond Borrowed Histories: Person-Aligned User Simulation for Interactive Role-Playing Evaluation","summary":"Role-playing agents (RPAs) have become one of the most important consumer applications of large language models. Users engage in multi-turn conversations with RPAs for experiences such as emotional comfort, making reliable evaluation essential for measuring capability, comparing systems, and guiding further improvement. Existing benchmarks, however, typically require an RPA to continue a fixed dialogue history and then evaluate the continuation using a fixed rubric detached from the user. We identify and empirically demonstrate two limitations of this design. First, an RPA's output is shaped by the preceding dialogue history, preventing a scientifically grounded assessment of its role-playing ability in real multi-turn settings. Second, user experience varies substantially across individu…","authors":["Yuhang Zhu","Mingxuan Du","Benfeng Xu","Jie Gao","Lingyun Yu","Hongtao Xie"],"publishedAt":"2026-07-30T00:00:00.000Z","submittedAt":null,"upvotes":30,"num_comments":2,"thumbnail":"https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2607.27816.png","hf_url":"https://huggingface.co/papers/2607.27816","arxiv_url":"https://arxiv.org/abs/2607.27816","github_repo":null,"github_stars":null,"ai_keywords":[]},{"paperId":"2607.28509","title":"RefCaptioner: Multi-Reference Image-Grounded Video Captioning","summary":"Existing video captioning models generate natural descriptions of video content but cannot explicitly ground local visual elements to multiple reference images. We introduce multi-reference image-grounded video captioning, a new task requiring factual video descriptions with phrase-level reference grounding, and propose RefCaptioner, a two-stage post-training framework for this task. RefCaptioner combines mixed-data SFT with Hierarchical Coverage-Discounted GRPO to jointly improve reference selection, phrase-level binding, distractor rejection, and cross-reference consistency while preserving general video-captioning ability. To support training, we construct a corpus containing 20,000 videos and 171,354 reference images. We further introduce MRVBench, a benchmark for evaluating caption f…","authors":["Tengfei Liu","Yang Shi","Yuran Wang","Xiaohan Zhang","Yuqing Wen","Yuqi Tang","Qixun Wang","Zhuoran Zhang"],"publishedAt":"2026-07-30T00:00:00.000Z","submittedAt":null,"upvotes":25,"num_comments":2,"thumbnail":"https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2607.28509.png","hf_url":"https://huggingface.co/papers/2607.28509","arxiv_url":"https://arxiv.org/abs/2607.28509","github_repo":null,"github_stars":null,"ai_keywords":[]},{"paperId":"2607.27146","title":"MindForge: Teaching Small Language Models Whole-Life-Cycle Software Engineering via Source-Free Program Synthesis","summary":"Coding agents have made substantial progress on software engineering tasks that modify existing codebases, including bug fixing and feature implementation. However, constructing a complete program from scratch remains a major challenge: even the frontier models evaluated on ProgramBench fully resolve fewer than 1% of tasks. One obstacle is the lack of scalable training environments for this from-scratch setting, spanning the whole software engineering life cycle, as existing environment-construction frameworks focus only on a single phase in software development. To address this gap, we introduce MindForge, an automated pipeline that converts open-source command-line programs into source-free environments that expose only a compiled reference executable and its documentation. Using MindFo…","authors":["Yihao Chen","Shi Chang","Khaled Chawa","Feng Lin","Boyuan Chen","Shaowei Wang","Ahmed E. Hassan"],"publishedAt":"2026-07-29T00:00:00.000Z","submittedAt":null,"upvotes":23,"num_comments":3,"thumbnail":"https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2607.27146.png","hf_url":"https://huggingface.co/papers/2607.27146","arxiv_url":"https://arxiv.org/abs/2607.27146","github_repo":null,"github_stars":null,"ai_keywords":[]},{"paperId":"2607.26769","title":"See2Think: Do Multimodal Models Really Use Intermediate Visual States?","summary":"Multimodal large language models increasingly use sketches, annotations, tools, and intermediate images during reasoning, but it remains unclear whether they truly rely on these visual states. Existing benchmarks are limited both by task collections with narrow coverage or partially text-solvable samples and by evaluations that emphasize final answers without diagnosing how intermediate visual states are generated, rendered, and used. We introduce See2Think, a unified evaluation framework comprising See2ThinkBench and Visual Action-of-Thought (VAoT). See2ThinkBench contains 1,200 open-ended, visually dependent problems across 12 task categories spanning 2D structured, 3D scene, and real-world reasoning. VAoT records textual thoughts, visual actions, rendered states, and subsequent reasoni…","authors":["Siyu Yan","Zhuoran Yan","Haiying Xu","Panhao Zhou","Jingyu Chen","Chenhao Ji","Shuo Cao","Yongheng Zhang"],"publishedAt":"2026-07-29T00:00:00.000Z","submittedAt":null,"upvotes":22,"num_comments":2,"thumbnail":"https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2607.26769.png","hf_url":"https://huggingface.co/papers/2607.26769","arxiv_url":"https://arxiv.org/abs/2607.26769","github_repo":null,"github_stars":null,"ai_keywords":[]},{"paperId":"2607.27703","title":"SpatialCLI: Learning to Reason With Spatial Tools, Then Without Them","summary":"Vision-language models (VLMs) are increasingly used in embodied agents to interpret visual inputs, reason about spatial relationships, and make task-level decisions based on that reasoning. However, a fundamental capability mismatch remains: general VLMs can reason about the overall task but often miss the visual details that determine success, while specialist vision models can capture those details but cannot translate them into task-level decisions. In this work, we propose SpatialCLI, a framework that teaches VLMs to reason with spatial tools and progressively internalize the specialist perceptual capabilities they provide. SpatialCLI proceeds in three stages: (1) Call exposes specialist vision models as spatial tools to augment the VLM's perception; (2) Learn uses Cold-Start SFT and …","authors":["Yang Zhou","Zixuan Huang","Sunzhu Li","Zhuo Yang","Chen Zhang","Shunian Chen","Caijun Yan","Jianyao Xu"],"publishedAt":"2026-07-30T00:00:00.000Z","submittedAt":null,"upvotes":21,"num_comments":2,"thumbnail":"https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2607.27703.png","hf_url":"https://huggingface.co/papers/2607.27703","arxiv_url":"https://arxiv.org/abs/2607.27703","github_repo":null,"github_stars":null,"ai_keywords":[]},{"paperId":"2607.27167","title":"SpecFirst: Behavioral Specification Elicitation as a First-Class Step in Agent-Based Program Synthesis from Scratch","summary":"LLM-based agents excel at software engineering tasks where an existing codebase provides context, but constructing a program from scratch remains fundamentally harder. Recent benchmarks such as ProgramBench quantify this gap: given only natural-language documentation and an execute-only binary as a behavioral oracle, even frontier models solve fewer than 1% of instances. Existing frameworks conflate documentation reading, behavioral exploration, and code synthesis into a single pass, causing agents to probe insufficiently, lose behavioral intent as context drifts, and propagate early misinterpretations into the final implementation. Inspired by classical requirements engineering, we argue that behavioral specification elicitation should be a first-class phase that precedes implementation.…","authors":["Yihao Chen","Shi Chang","Feng Lin","Khaled Chawa","Boyuan Chen","Shaowei Wang","Ahmed E. Hassan"],"publishedAt":"2026-07-29T00:00:00.000Z","submittedAt":null,"upvotes":18,"num_comments":2,"thumbnail":"https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2607.27167.png","hf_url":"https://huggingface.co/papers/2607.27167","arxiv_url":"https://arxiv.org/abs/2607.27167","github_repo":null,"github_stars":null,"ai_keywords":[]},{"paperId":"2607.28582","title":"β-OPSD: Deriving with Policy Optimization, Training with Self-Distillation","summary":"On-policy self-distillation (OPSD) is a promising approach to improve reasoning language models, but it remains brittle in practice: making it work reliably often requires substantial engineering effort. We identify a structural source of this difficulty: vanilla OPSD is precisely the β=1 member of a broader policy-optimization family, where β weights the KL penalty anchoring the student to a reference policy. This equivalence turns β from an implicit value fixed at one into a controllable regularization parameter, yielding a more general formulation that trades off proximity to a reference policy against privileged teacher guidance. We introduce β-OPSD and derive its optimal policy as a geometric interpolation between the reference policy and the privileged teacher. Directly optimizing t…","authors":["Jiawei Xu","Minghui Liu","Juzheng Zhang","Tom Goldstein","Furong Huang"],"publishedAt":"2026-07-30T00:00:00.000Z","submittedAt":null,"upvotes":18,"num_comments":2,"thumbnail":"https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2607.28582.png","hf_url":"https://huggingface.co/papers/2607.28582","arxiv_url":"https://arxiv.org/abs/2607.28582","github_repo":null,"github_stars":null,"ai_keywords":[]},{"paperId":"2607.28362","title":"ShadowDancer: Teaching Video World Models Any Action by Learning Unified Dynamics Representations from a Video and Its Shadow","summary":"We present ShadowDancer, a novel approach to any-action, frame-level control of interactive video world models. The obstacle is representational: existing interfaces either encode an action loosely, leaving how it unfolds for the model to improvise, or encode it exactly through structured signals that serve one family and are hard to acquire, so precise control across diverse dynamics remains impractical. Demonstration videos are the natural remedy, specifying any dynamics frame by frame; yet a video shows its dynamics only through one particular appearance, a single shadow of the underlying dynamics, so actions learned from demonstrations transfer poorly to new scenes. ShadowDancer addresses this with two key innovations: (1) shadow pairs, video pairs that replay the same dynamics under …","authors":["Jin Cao","Zian Meng","Kaipeng Zhang"],"publishedAt":"2026-07-30T00:00:00.000Z","submittedAt":null,"upvotes":16,"num_comments":3,"thumbnail":"https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2607.28362.png","hf_url":"https://huggingface.co/papers/2607.28362","arxiv_url":"https://arxiv.org/abs/2607.28362","github_repo":null,"github_stars":null,"ai_keywords":[]},{"paperId":"2607.26754","title":"StatePlay: State-Aware Game World Models for Mechanics-Consistent Generation","summary":"Recent game world models can generate visually realistic and interactive environments conditioned on player actions. However, games are not defined by pixels alone; they are governed by explicit mechanics, namely state-dependent rules that control health reduction, skill activation, and game termination. These mechanics depend on precise internal states, such as health points, skill meters, and timers, which are tightly coupled with visual observations and determine how gameplay evolves. Without modeling these state dynamics, existing game world models may generate visually plausible rollouts but violate the underlying game rules. In this paper, we propose StatePlay, a novel state-aware game world model that jointly predicts visual content and game states to promote mechanics-consistent g…","authors":["Zijun Lin","Zeqing Wang","Cheston Tan","Bihan Wen","Yeying Jin"],"publishedAt":"2026-07-29T00:00:00.000Z","submittedAt":null,"upvotes":16,"num_comments":2,"thumbnail":"https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2607.26754.png","hf_url":"https://huggingface.co/papers/2607.26754","arxiv_url":"https://arxiv.org/abs/2607.26754","github_repo":null,"github_stars":null,"ai_keywords":[]},{"paperId":"2607.28611","title":"Chimera: Designing and Chinchilla-Scaling Hybrid Visual Diffusion Transformers","summary":"Visual generation increasingly requires high-resolution images, long videos, and multimodal context, making the quadratic cost of full attention prohibitive. We introduce Chimera, a hybrid visual diffusion backbone with a principled scaling recipe. Chimera processes text, image, and video tokens in one raster-ordered stream without positional embeddings. It combines Kimi Delta Attention (KDA) for long-context state tracking with O(N) complexity, interleaved Multi-head Latent Attention (MLA) for direct global interaction, and modality-aware short convolutions for local spatiotemporal context. Sparse Mixture-of-Experts (MoE) layers expand capacity while controlling activated compute. To scale this heterogeneous architecture, we introduce HeteroP, a module-wise scheme that transfers hyperpar…","authors":["Chongjian Ge","Hanwen Jiang","Tianyu Wang","Jiuxiang Gu","Yiran Xu","Ziwen Chen","Shaoteng Liu","Jing Shi"],"publishedAt":"2026-07-30T00:00:00.000Z","submittedAt":null,"upvotes":16,"num_comments":1,"thumbnail":"https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2607.28611.png","hf_url":"https://huggingface.co/papers/2607.28611","arxiv_url":"https://arxiv.org/abs/2607.28611","github_repo":null,"github_stars":null,"ai_keywords":[]},{"paperId":"2607.28410","title":"Can Large Language Models Execute Parent Orders?","summary":"Parent-order execution is a core problem in algorithmic trading, where the goal is to split a large order into smaller orders while reducing execution costs. Existing approaches either rely on pre-specified market assumptions that may not hold in practice, or require task-specific training that limits adaptability to new settings. To overcome these limitations, we present the first systematic study of large language models (LLMs) for parent-order execution. This extends the use of LLMs in finance from what to trade to how to execute. We propose PACE (Plan-Ahead Controlled Execution), a hierarchical framework that decomposes parent-order execution into long-horizon planning and short-horizon execution, requiring neither explicit market assumptions nor task-specific training. Experiments on…","authors":["Zane Shen","Xinli Xu","Guangyi Zhang","Jialong Chen","Jinsong Zhou","Cong Chen","Guibao Shen","Dongyu Yan"],"publishedAt":"2026-07-30T00:00:00.000Z","submittedAt":null,"upvotes":14,"num_comments":2,"thumbnail":"https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2607.28410.png","hf_url":"https://huggingface.co/papers/2607.28410","arxiv_url":"https://arxiv.org/abs/2607.28410","github_repo":null,"github_stars":null,"ai_keywords":[]},{"paperId":"2607.26056","title":"INTACT: Isomorphic Intent-to-Action Learning for Search-Free World Models","summary":"Forward latent world models predict how actions change a scene, but recover actions for a desired change only through expensive test-time search. We introduce INTACT (INtent-To-ACTion), an end-to-end JEPA that turns action-labeled, reward-free trajectories into a deployable intent-to-action interface. Each transition supplies physical intent z_{t+1}-z_t, while a future goal supplies deployment intent sg(z_g)-z_t. The architecture is isomorphic between the local and goal motion-intent backbone-input graphs through an identical four-slot grammar and shared parameters, and between supported local and goal motion-intent families through action-law semantics induced by the same predictor rather than pointwise latent equality. INTACT also provides intact transfer from RGB evidence to action-eff…","authors":["Junhan Sun","Hao Zhao","Guofeng Zhang"],"publishedAt":"2026-07-28T00:00:00.000Z","submittedAt":null,"upvotes":13,"num_comments":3,"thumbnail":"https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2607.26056.png","hf_url":"https://huggingface.co/papers/2607.26056","arxiv_url":"https://arxiv.org/abs/2607.26056","github_repo":null,"github_stars":null,"ai_keywords":[]},{"paperId":"2607.28272","title":"MemHarness: Memory Is Reconstructed, Not Replayed","summary":"Retrieving past experiences has become a common strategy to enhance large language model agents. However, most existing memory-augmented agents treat retrieved experiences as static records to be replayed verbatim, injecting them into the context regardless of whether they align with the agent's current situation. This ``replay'' paradigm ignores the gap between the abstract, general nature of stored experience and the concrete, ever-changing states encountered at decision time, frequently causing negative transfer. In contrast, humans rarely recall past experiences verbatim; instead, they reorganize and adapt retrieved memories to fit the present context. Inspired by this, we propose MemHarness, a framework that equips LLM agents to actively harness and reconstruct past experiences based…","authors":["Rong Wu","Daocheng Fu","Licheng Wen","Xuemeng Yang","Shu Zou","Jianbiao Mei","Yuxin Wang","Hairong Zhang"],"publishedAt":"2026-07-30T00:00:00.000Z","submittedAt":null,"upvotes":13,"num_comments":2,"thumbnail":"https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2607.28272.png","hf_url":"https://huggingface.co/papers/2607.28272","arxiv_url":"https://arxiv.org/abs/2607.28272","github_repo":null,"github_stars":null,"ai_keywords":[]},{"paperId":"2607.27372","title":"Explorative Modeling: Unlocking a Third Pretraining Axis and End-to-End Generation","summary":"The deep learning revolution, kicked off by AlexNet, taught us that end-to-end training beats decomposing a problem into hand-designed stages. Generative modeling, however, has remained the exception-despite generative models being remarkably capable, they are still not trained end-to-end. This is because, at its core, generative modeling is about handling distributions with many modes, and existing scalable approaches handle this the same way, by factoring the generation procedure, which prevents end-to-end generation. In this work, we introduce Explorative Modeling, a new paradigm that instead factors the training loop, exploring K candidate matches between model generations and data, and training on the best, so predictions commit to modes rather than blurring them. We find Explorative…","authors":["Alexi Gladstone","Heng Ji","Yilun Du"],"publishedAt":"2026-07-29T00:00:00.000Z","submittedAt":null,"upvotes":12,"num_comments":2,"thumbnail":"https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2607.27372.png","hf_url":"https://huggingface.co/papers/2607.27372","arxiv_url":"https://arxiv.org/abs/2607.27372","github_repo":null,"github_stars":null,"ai_keywords":[]},{"paperId":"2607.27278","title":"OVEarth-Bench: Evaluating Category Breadth and Query Diversity for Open-Vocabulary Earth Observation","summary":"Open-vocabulary Earth observation (EO) aims to localize geospatial concepts specified in natural language rather than a fixed label set. Existing benchmarks, however, usually cover narrow category vocabularies or limited query forms. To fill this gap, we introduce OVEarth-Bench, which extends existing evaluation in two directions: category breadth, through broad hierarchical category coverage with positive and negative expressions, and query diversity, through vocabulary, referring, and reasoning queries. The benchmark supports mask and box localization under a unified zero-shot protocol. We evaluate a broad set of general and EO-specific methods. The evaluation reveals that: (1) the performance of current methods remains limited, while broader category coverage yields more stable model r…","authors":["Kaiyu Li","Zepeng Xin","Zixuan Jiang","Jing Fu","Lanxuan Xue","Lingyu Zhang","Xiangyong Cao"],"publishedAt":"2026-07-29T00:00:00.000Z","submittedAt":null,"upvotes":12,"num_comments":2,"thumbnail":"https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2607.27278.png","hf_url":"https://huggingface.co/papers/2607.27278","arxiv_url":"https://arxiv.org/abs/2607.27278","github_repo":null,"github_stars":null,"ai_keywords":[]},{"paperId":"2607.27958","title":"Σ-Mem: An Online Reliability Memory for LLM-based Multi-Agent Systems","summary":"Memory is central to long-horizon LLM agents, yet existing memory systems primarily preserve interaction content rather than modeling which agents can be trusted and under what conditions. This limitation is particularly important in multi-agent systems, where a central model may be unable to directly verify plausible or correlated peer responses. We introduce Σ-Mem, an online reliability memory that records historical competence evidence for individual peers and peer relationship evidence across the peer set. Both forms of evidence are maintained as real symmetric states and updated from post-decision correctness feedback. By Weyl's inequality, the spectral change caused by each event-level update is bounded, enabling stable online adaptation without retraining the underlying models. Σ-M…","authors":["Peilin Feng","Suorong Yang","Soujanya Poria"],"publishedAt":"2026-07-30T00:00:00.000Z","submittedAt":null,"upvotes":12,"num_comments":2,"thumbnail":"https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2607.27958.png","hf_url":"https://huggingface.co/papers/2607.27958","arxiv_url":"https://arxiv.org/abs/2607.27958","github_repo":null,"github_stars":null,"ai_keywords":[]}],"summary":{"by_keyword":[],"most_upvoted":{"paperId":"2607.28618","title":"AskChem: Claim-Centered Infrastructure for Chemistry Literature Synthesis","upvotes":292},"most_discussed":{"paperId":"2607.28227","title":"Qwen-UI-Agent Technical Report: Toward Next-Generation Real-World Centric Foundation GUI Agents","comments":7}}}}