🏷️ Research Paper
Real-time tracking of AI industry hot topics · 56 articles
WSJ：三名研究人员用 Claude Opus 5 借 Discourse 漏洞访问 OpenAI 私有代码
据 WSJ 报道，三名研究人员使用 Claude Opus 5 将 Discourse 的一个漏洞串联成对 OpenAI 私有代码的访问。 researchers 在 Discourse 服务器上获得了包括 OpenAI 员工在内的认证 token，部分 forum token 可用于 ChatGPT 并触达 OpenAI 的 GitHub 服务。 🔗 阅读原文 via AIHOT · https://aihot.news/items/cmu6i7irl000qro0fik1cusru
0 Read originalAnthropic 用 Claude 优化 30 多个开源生物分子模型，平均提速约 4 倍并开源全部代码
Anthropic 发布研究，让 Claude 在不到四周内优化了 30 多个开源生物分子模型，平均提速约 4 倍，输出完全一致时约 2 倍。 🔗 阅读原文 via AIHOT · https://aihot.news/items/cmu5y1dkt069qroiqnnhxz778
0 Read original论文：Coding Agents with an Obstacle-Aware Harness for Safe Robot Manipulation
Coding agents have emerged as a promising paradigm for robot manipulation: a language model writes the robot controller as a program, and agents built in this way now operate robots without robot-specific training.Whether this paradigm is also safe, however, has not been asked. We evaluate coding agent under a safety constraint, where each task pairs a manipulation goal with an obstacle the robot
0 Read original论文：Embedding Models Measure in Peculiar Ways
Embedding spaces define notions of semantic similarity and distance. We study whether those embeddings reflect physical measurements of mass, distance, time and volume, which admit a unique, objective notion of semantic equivalence and distance. We find that physical measurement is only weakly modeled in the embedding space, and that instead quite peculiar measurement patterns can be observed. Fur
0 Read original论文：Workspace Models: Lightweight Robotic Memory via Saliency-Driven Supervision
Complex robotic manipulation tasks frequently require a long-term memory of past events and actions. As conditioning on full histories renders policies prone to spurious correlations and degrades performance, many approaches to policy memory involve compressing historical information through expensive VLM queries in-the-loop to process only task-salient information. In this paper, we propose an al
0 Read original论文：FAMOS: Feed-Forward 3D Articulation Modeling from Sparse Observations
Modeling articulated objects from sparse monocular views is challenging because each observation reveals only partial geometry and motion evidence. Most feed-forward methods infer articulation from a single observation and therefore rely heavily on learned category-level shape priors. We present FAMOS, a feed-forward model that predicts movable-part segmentation and joint parameters from a sparse,
0 Read original论文：Paint-Anything: Unified Any-Color Control for Image Generation and Editing
Professional design requires any-color control: the ability to specify an object's target color with any 24-bit hex value for image generation and editing. Prior work has explored color generation, editing, and colorization, but often relies on dedicated color representations or specialized inference procedures. Advances in large language models offer a simpler starting point: even compact models
0 Read original论文：ERCPMP-Gx: Endoscopic Image and Video Dataset for Morphological, Histopathological, and Genomic Characterization of Colorectal Polyposis
Hereditary polyposis syndromes can be precursor lesions to colorectal cancer and are associated with a broad spectrum of extracolonic tumors. Early identification and accurate classification of these syndromes are essential for timely diagnosis, individualized patient management, and targeted surveillance strategies for affected families. However, public endoscopic datasets are largely organized a
0 Read original论文：How Does Distribution Shift Shape Pretraining Gains in Neural PDE Surrogates?
Pretraining a neural PDE surrogate can reduce the amount of new CFD data needed when geometry or modeled physics changes. However, it remains unclear how different components of distribution shift affect this benefit. We pretrain a surrogate on 254,909 RANS solutions from one airfoil family and fine-tune it on a new family under two target settings with matched freestream ranges: the same Spalart-
0 Read original论文：Quantifying Overclaiming Propensity in Frontier LLM Agents
Frontier coding agents are increasingly trusted to work autonomously for long periods, yet an agent's final response is often the only account of that work a user sees. We quantify the propensity of frontier agents to \emph{overclaim} task completion, a misrepresentation that can mislead the user. An agent overclaims when its final response contradicts information in its context. This definition r
0 Read original论文：Unifying Models of Intergroup Hostility in Online Discourse
Hostile rhetoric toward social groups can normalize exclusion and justify mistreatment, as well as contribute to rising polarization and political violence. Efforts to moderate hostile rhetoric in online speech draw on foundational theories in social and moral psychology, and political science. However, these theories were developed largely in parallel, often propose different and sometimes confli
0 Read original论文：Score Centering Stabilizes Off-policy Reinforcement Learning
Reinforcement learning (RL) of large language models is notoriously sensitive to small differences between training and inference engines, often referred to as the training-inference mismatch (TIM). However, completely eliminating TIM is impractical, as it would come at a major cost to rollout efficiency. In this paper, we show that the instability of RL under TIM is primarily caused by drift: a p
0 Read original论文：An Empirical Study of Harness Design for Coding Agents
Coding harnesses shape how autonomous coding agents translate model capabilities into long-horizon software-engineering performance, yet existing work typically evaluates harnesses as monolithic systems, leaving the effectiveness of individual components unclear. To enable component-level comparisons, we study this question with a lightweight coding harness whose execution loop is fixed while thre
0 Read original论文：JEPA-Anything: Learning Predictive Models across Different Worlds
World modeling enables intelligence to anticipate consequences, guide interventions, and learn from interaction. Yet predictive models remain domain-specific: can a common learning principle support world modeling across radically different systems? We introduce JEPA-Anything, a domain-agnostic framework based on orthogonal predictive factorization (OPF). Extending joint-embedding predictive archi
0 Read original论文：PosteriorBench: From Point Estimates to Posterior Matching in Evaluating Generative Inverse Solvers
Generative models are increasingly used to solve scientific inverse problems, but existing evaluations still focus primarily on whether a method can produce a single plausible reconstruction. This is insufficient for ill-posed problems, where multiple solutions may be consistent with the same sparse or noisy observations. In these settings, a method can achieve strong pointwise accuracy while stil
0 Read original论文：RetireOPD: Self-Retiring On-Policy Distillation for Agentic Reinforcement Learning
Multi-turn agents trained with reinforcement learning (RL) receive a single scalar reward per trajectory, which motivates self on-policy distillation (OPD) to supply dense token-level supervision from a self-teacher with privileged task skills, letting a skill-free student internalize them. This recipe, however, is undermined by two findings in agentic tasks: privileged information alone does not
0 Read original论文：Harm Laundering in GPT Models: Evidence That Gender Discrimination Is Transformed Rather Than Reduced Across Safety-Trained Generations
Safety evaluations for large language models rely on surface-form classifiers that report declining harm scores across model generations. We provide evidence that this methodology is systematically incomplete: explicit discriminatory content is transformed rather than removed. We call this \emph{harm laundering}. Analysing 450,000 gender-directed completions across 15 models spanning GPT-2 through
0 Read original论文：GeoAAC: Geometry-Based Adaptive Action Chunking from Denoising Trajectories in VLA Policies
Action chunking is widely used for action generation and execution in Vision-Language-Action (VLA) policies, yet existing approaches commonly use a fixed action horizon. During a rollout, different task stages may require different levels of action continuity, control precision, and closed-loop feedback, making a fixed horizon unable to accommodate changing control requirements. We propose \textbf
0 Read original论文：Semantic Action Graph: A Shared Representation for Agent Grounding and Human Interpretation of Sports Highlights
Generative agents are increasingly used to select and narrate video highlights, but they typically operate over unstructured or frame-level representations. Their output is consequently difficult for a viewer to verify and steer toward individual preferences. We present the semantic action graph, a lightweight domain schema that represents a sports match as performer, action, recipient, moment, an
0 Read original论文：Calibrated RF-Fingerprinting Under Interference With Heterogeneous Transmission Protocols
Radio Frequency(RF)-Fingerprinting is a spectrum monitoring technique that identifies specific transmitters based on hardware impairments imprinted within the emitted signal. Although widely researched, studies almost exclusively consider scenarios where only one transmitter is emitting at a time, limiting real world applicability. In this work, we further the study of RF-Fingerprinting by conside
0 Read original论文：Agile-WAM: An Agile Tactile World Action Model for Contact-Rich Robot Control
World Action Models (WAMs) advance beyond conventional visuomotor policies by jointly predicting future world states and robot actions, enabling the policy to learn physical dynamics that support effective control. However, recent tactile WAMs often rely on large-scale pretrained generative backbones to capture contact-rich physical dynamics, which limit their inference efficiency and flexible dep
0 Read original论文：Prediction-Powered Smoothing and Validation for Disaggregated AI Evaluation
Evaluating an AI system requires disaggregated assessment, as performance varies across domains such as benchmark task types or conversation types in deployed agents. Exhaustive testing is expensive, so evaluation rests on a sample of labeled units. We treat the evaluation set as a finite population and seek accurate point and interval estimates of each domain mean. Direct estimators, including pr
0 Read original论文：OPTED: On-Policy Fine-Tuning for End-to-End Driving using a Render-Free Teacher
As scaling pre-training data alone yields diminishing returns, post-training is becoming increasingly important across physical AI domains such as autonomous driving. End-to-end driving policies are pre-trained in open loop with behavior cloning on human demonstrations. However, compounding errors during closed-loop deployment can take the vehicle outside the training data distribution, increasing
0 Read original论文：RAFT: A Stateful Retrieval-Augmented Framework for Troubleshooting Agents
Effective troubleshooting agents in enterprise customer support depend on retrieving actionable guidance from similar historical cases, yet existing retrieval-augmented generation (RAG) systems treat support cases as static documents and overlook their multi-stage, stateful nature. We introduce RAFT (Retrieval-Augmented Framework for Troubleshooting Agents), a stateful RAG framework that abstracts
0 Read original论文：Large Language Models as Falsifiers for Cyber-Physical Systems
Falsification searches for counterexamples to formal specifications in cyber-physical systems (CPS). With specifications written in Signal Temporal Logic (STL), falsification can be formulated as a robustness optimization problem, traditionally tackled with black-box search algorithms. In parallel, large language models (LLMs) have recently emerged as surprisingly effective optimizers when coupled
0 Read original论文：dQwen3.5: Hybrid-Attention Diffusion Language Models
Adapting a pretrained autoregressive (AR) model is a cost-efficient route to a diffusion language model (DLM). While nearly all such adaptations start from a full-attention transformer, AR modeling has shifted toward hybrid architectures that interleave attention and RNN layers. This creates an obstacle for adaptation: unlike attention, RNNs are structurally causal and nontrivial to bidirectionali
0 Read original论文：MILER: Semantic Mid-Level Representation for Sim-to-Real Reinforcement Learning in Unstructured Autonomous Driving
Reinforcement learning constitutes a promising approach owing to its potential for superhuman performance and self-learned policies. However, its application to real-world autonomous driving remains scarce, particularly in unstructured environments, because of the challenges associated with sim-to-real transfer for unstructured environments. In this work, we present MILER, an end-to-end policy fra
0 Read originalGoodfire Research 发现模型内部信号可规模化检测奖励作弊
Goodfire Research 发现模型内部存在伴随奖励作弊的激活信号，可用简单探针实时检测。在 Kimi K3、GLM 5.2、Qwen 3.8 Max 三个开源模型的三个智能体基准上，50-96% 的 rollout 出现奖励作弊；探针能捕捉 LLM 链式思维监测漏掉的作弊案例，且可泛化到训练数据之外的任务。 🔗 阅读原文 via AIHOT · https://aihot.news/items/cmu5r7kl30h35roqonpk4qypn
0 Read originalDwarkesh 对谈 Noam Brown：智能体集群、对齐与递归自我改进
Dwarkesh Patel 采访 OpenAI 研究员 Noam Brown，谈多智能体系统、对齐与递归自我改进。 🔗 阅读原文 via AIHOT · https://aihot.news/items/cmu5qsx2y0gpfroqonmogw4sk
0 Read original论文：Region-Level Policy Optimization for Fine-grained MLLM Perception
Region-Level Policy Optimization for Fine-grained MLLM Perception
0 Read original论文：SoL-Pi: Recursively Scaling Auto-Research Loops for Efficient Agent Harness
SoL-Pi: Recursively Scaling Auto-Research Loops for Efficient Agent Harness
0 Read original论文：An Empirical Study of Harness Design for Coding Agents
An Empirical Study of Harness Design for Coding Agents
0 Read original论文：WeVisDoc: From Coverage to Capability for Robust End-to-End Document Parsing
WeVisDoc: From Coverage to Capability for Robust End-to-End Document Parsing
0 Read original论文：RetireOPD: Self-Retiring On-Policy Distillation for Agentic Reinforcement Learning
RetireOPD: Self-Retiring On-Policy Distillation for Agentic Reinforcement Learning
0 Read original论文：When2Think: Learning Difficulty-Aware Length Control for Efficient Hybrid Reasoning Models
When2Think: Learning Difficulty-Aware Length Control for Efficient Hybrid Reasoning Models
0 Read original论文：What Does Privileged Information Add to On-Policy Self-Distillation?
What Does Privileged Information Add to On-Policy Self-Distillation?
0 Read original论文：DeepSeek-V4.1-Flash: Pushing the Limits of KV Cache Compression
DeepSeek-V4.1-Flash: Pushing the Limits of KV Cache Compression
0 Read original论文：UFO: Chain-of-Evaluation for Omni-Condition Alignment in Multi-Modal Image Generation
UFO: Chain-of-Evaluation for Omni-Condition Alignment in Multi-Modal Image Generation
0 Read original论文：Video DeltaNet: A Video-Native Hybrid Attention for Livestream Video Generation
Video DeltaNet: A Video-Native Hybrid Attention for Livestream Video Generation
0 Read original论文：JEPA-Anything: Learning Predictive Models across Different Worlds
JEPA-Anything: Learning Predictive Models across Different Worlds
0 Read original论文：VABench: Measuring Embodied Spatial Intelligence through Visual Demonstrations, Active Perception, and Metric Control
VABench: Measuring Embodied Spatial Intelligence through Visual Demonstrations, Active Perception, and Metric Control
0 Read original论文：FAMOS: Feed-Forward 3D Articulation Modeling from Sparse Observations
FAMOS: Feed-Forward 3D Articulation Modeling from Sparse Observations
0 Read original论文：When EOS Tokens Disagree: Understanding Length Inflation in On-Policy Distillation
When EOS Tokens Disagree: Understanding Length Inflation in On-Policy Distillation
0 Read original论文：VākQA: A Benchmark and Evaluation Study for Telugu Spoken Factoid Question Answering
VākQA: A Benchmark and Evaluation Study for Telugu Spoken Factoid Question Answering
0 Read original论文：Don't Mask the Environment: Observation Supervision Changes How Agents Explore Under RL
Don't Mask the Environment: Observation Supervision Changes How Agents Explore Under RL
0 Read original论文：Objective vs. Search: Decomposing What Makes a Good Tokeniser
Two dominant tokenisation algorithms are used by modern language models: byte-pair encoding (BPE) and UnigramLM. These differ along two orthogonal axes: their optimisation objective (compression vs. log-likelihood) and their search procedure (bottom-up merging vs. top-down pruning). Existing comparisons confound these axes, making it unclear whether their observed differences stem from what is bei
0 Read original论文：A Zeroth-Order Paradigm for LLM Preference Alignment
Direct preference alignment methods are widely used to align large language models (LLMs) with human preferences because of their computational and memory efficiency. However, likelihood displacement motivates alternative ways to extract information from preference pairs with small likelihood margins. In this paper, we propose and analyze Comparison-based Preference Optimization (ComPO), a zeroth-
0 Read original论文：PANORAMA: Panoptic Grounded Captioning via Mask Proposal Selection
Intelligent systems that act in the world require image understanding that is both comprehensive and spatially grounded. Current vision-language models (VLMs) can generate fluent and detailed image captions, but reliably associating them with image pixels remains challenging. Existing methods that combine dense captioning with pixel-level grounding often produce either incomplete descriptions or i
0 Read original论文：Dreaming the Sound of Contact: Leveraging Video and Audio Generation for Zero-Shot Force-Aware Manipulation and Data Generation
Recent advances in video generation allow robots to learn manipulation trajectories from generated videos. However, these approaches produce purely kinematic trajectories that lack force information, causing failures in contact-rich tasks where appropriate contact forces are essential for success. In this work, we explore augmenting generated video with audio to shape a bounded, time-varying desir
0 Read original论文：Exponential Hardness of Off-Policy Evaluation under History-Dependent Logging
Can a logged dataset visit every hidden state frequently and still be exponentially uninformative about a target policy's value? We show that it can when the logger depends on history. For every horizon $H \ge 3$, we construct two POMDPs with at most two latent states per stage, three actions, and a common logger with three memory states. Action coverage, belief coverage, and two behavior-marginal
0 Read original论文：ScienceIDE: Turning World's Scientific Codebase into Agent Learnable Environments
Scientific code repositories encode decades of human knowledge in executable models, methods, and tools. Yet fragmented toolchains, implicit domain conventions, and specialized correctness criteria make this knowledge difficult to convert into reliable learning experience-a challenge we call the scientific experience bottleneck. We introduce ScienceIDE, infrastructure for turning the world's scien
0 Read original论文：Cognitive Extensions for Dual-Process Language Agents: Memory and Self-Reflection in Interactive Environments
Language agents remain brittle in interactive environments, where success requires long-horizon state tracking, valid action execution, and recovery from failed steps. We extend SwiftSage, a dual-process agent that combines a fast action proposer with a slower planner, using two modular cognitive extensions: an Adaptive Memory Module (AMM) for salience-gated episodic storage and trigger-driven ret
0 Read original论文：Affora: A Design System for Agent-Friendly Interfaces
Computer-use agents increasingly operate software designed for people, but interfaces often leave actions or task state unclear to machine readers. We present Affora, a design system that supports both readers while preserving visual freedom and familiar human workflows. Three controlled studies examine component implementations, visual variation, and interaction-design principles. Their findings
0 Read original论文：Flag Game: A Toy Model for Mechanistic Swarm Interpretability
Emergent coordinated behaviors of AI agents are starting to present critical safety risks. A key phenomenon driving these behaviors is the rapid formation and spread of beliefs about the world, and mechanistic understanding is crucial for collective alignment. To this end, we introduce the Flag Game, a toy model for studying the mechanisms of collective belief formation. Concretely, a hidden count
0 Read original论文：Playing log(N)-Questions over Wikipedia Abstracts: Communication Efficiency Between Paired Frontier Models
We evaluate six frontier language models on the two-agent $\log(N)$-Questions game. A questioner sees $N$ Wikipedia lead paragraphs and must identify a secretly chosen target using exactly $\log_2 N$ yes/no questions. An answerer sees only the target and the question, and replies with one word. Both roles run on the same provider, so the game measures how well a model communicates with itself acro
0 Read original论文：How Model Growth, Recursion, and Boundary Operators Influence Scaling Exponents
Scaling laws predict how loss decreases with increases in computation. We show, contrary to conventional wisdom, that architectural interventions can modify scaling exponents in pre-training, leading to exponential improvements in performance with increases in computation. As an anchoring point, we consider the architectural formulation of looped transformers. Although not typically used in this w
0 Read original