探索全球前沿学术脉络

01.

arXiv (CS.CV) 2026-06-17 DOI: arXiv:2504.14582

NTIRE 2025 Challenge on Image Super-Resolution (x4): Methods and Results

作者:

Zheng Chen↗Kai Liu↗Jue Gong↗Jingkai Wang↗Lei Sun↗Zongwei Wu↗Radu Timofte↗Yulun Zhang↗Xiangyu Kong↗Xiaoxuan Yu↗Hyunhee Park↗Suejin Han↗…

This paper presents the NTIRE 2025 image super-resolution ($\times$4) challenge, one of the associated competitions of the 10th NTIRE Workshop at CVPR 2025. The challenge aims to recover high-resolution (HR) images from low-resolution (LR) counterparts generated through bicubic downsampling with a $\times$4 scaling factor. The objective is to develop effective network designs or solutions that achieve state-of-the-art SR performance. To reflect the dual objectives of image SR research, the challenge includes two sub-tracks: (1) a restoration track, emphasizes pixel-wise accuracy and ranks submissions based on PSNR; (2) a perceptual track, focuses on visual realism and ranks results by a perceptual score. A total of 286 participants registered for the competition, with 25 teams submitting valid entries. This report summarizes the challenge design, datasets, evaluation protocol, the main results, and methods of each team. The challenge serves as a benchmark to advance the state of the art and foster progress in image SR.

阅读与讨论 → 访问原文 →

02.

arXiv (CS.AI) 2026-06-18 DOI: arXiv:2603.00656

InfoPO: Information-Driven Policy Optimization for User-Centric Agents

作者:

Fanqi Kong↗Jiayi Zhang↗Mingyi Deng↗Chenglin Wu↗Yuyu Luo↗Bang Liu↗

arXiv:2603.00656v2 Announce Type: replace Abstract: Real-world user requests to LLM agents are often underspecified. Agents must interact to acquire missing information and make correct downstream decisions. However, current multi-turn GRPO-based methods often rely on trajectory-level reward computation, which leads to credit assignment problems and insufficient advantage signals within rollout groups. A feasible approach is to identify valuable interaction turns at a fine granularity to drive more targeted learning. To address this, we introduce InfoPO (Information-Driven Policy Optimization), which frames multi-turn interaction as a process of active uncertainty reduction and computes an information-gain reward that credits turns whose feedback measurably changes the agent's subsequent action distribution compared to a masked-feedback counterfactual. It then combines this signal with task outcomes via an adaptive variance-gated fusion to identify information importance while maintaining task-oriented goal direction. Across diverse tasks, including intent clarification, collaborative coding, and tool-augmented decision making, InfoPO consistently outperforms prompting and multi-turn RL baselines. It also demonstrates robustness under user simulator shifts and generalizes effectively to environment-interactive tasks. Overall, InfoPO provides a principled and scalable mechanism for optimizing complex agent-user collaboration. Code is available at https://github.com/kfq20/InfoPO.

阅读与讨论 → 访问原文 →

03.

arXiv (CS.LG) 2026-06-16 DOI: arXiv:2505.24275

GradPower: Powering Gradients for Faster Language Model Pre-Training

作者:

Jinbo Wang↗Mingze Wang↗Jiaqi Zhang↗Wei Wang↗Peng Pei↗Xunliang Cai↗Weinan E↗Lei Wu↗

arXiv:2505.24275v4 Announce Type: replace Abstract: We propose GradPower, a lightweight gradient-transformation technique for accelerating language model pre-training. Given a gradient vector $g=(g_i)_i$, GradPower first applies the elementwise sign-power transformation: $\varphi_p(g)=(sign(g_i)|g_i|^p)_{i}$ for a fixed $p>0$, and then feeds the transformed gradient into a base optimizer. Notably, GradPower requires only a single-line code change and no modifications to the base optimizer's internal logic, including the hyperparameters. When applied to Adam (termed AdamPower), GradPower consistently achieves lower terminal loss across diverse architectures (LLaMA, Qwen2MoE), parameter scales (66M to 2B), datasets (C4, OpenWebText), and learning-rate schedules (cosine, warmup-stable-decay). The most pronounced gains are observed when training modern mixture-of-experts models with warmup-stable-decay schedules. GradPower also integrates seamlessly with other state-of-the-art optimizers, such as Muon, yielding further improvements. Finally, we provide theoretical analyses that reveal the underlying mechanism of GradPower and highlight the influence of gradient noise.

阅读与讨论 → 访问原文 →

04.

arXiv (CS.LG) 2026-06-18 DOI: arXiv:2605.07022

Self-Driving Datasets: From 20 Million Papers to Nuanced Biomedical Knowledge at Scale

作者:

Haydn Jones↗Yimeng Zeng↗Alden Rose↗Li S. Yifei↗Yining Huang↗Kaiwen Wu↗Jiaming Liang↗Maggie Ziyu Huan↗Yoseph Barash↗Cesar de la Fuente-Nunez↗Osbert Bastani↗Zachary Ives↗…

arXiv:2605.07022v3 Announce Type: replace Abstract: Manually curated biomedical repositories – spanning bioactivity, genomics, and chemistry – are expensive to maintain, lag behind primary literature, and discard experimental context, obscuring nuances needed to assess data correctness and coverage. We show that PubMed itself can be autonomously and cost-effectively turned into structured datasets that are larger, more nuanced, and more accurate than the curated databases they replace. We present three coupled contributions: (1) an LLM-based entity-tagging pipeline, grounded in nine biomedical ontologies, that tags 4.5B entities across 19 categories in a 22.5M-paper, 2.5T-token PubMed corpus; (2) hybrid sparse-dense retrieval supporting entity-filtered semantic queries over the tagged corpus; and (3) Starling, a multi-agent deep research system that, given only a natural-language task description, designs precision- and recall-targeted retrieval filters, induces an extraction schema, and emits structured records with nuance-rich fields and supporting passages. Across six tasks – blood-brain barrier permeability, oral bioavailability, acute toxicity (LD50), gene-disease associations, protein subcellular localization, and chemical reactions – Starling produces ~6.3M records (91K-3M per task); several are, to our knowledge, the largest public datasets for their property. Frontier-model rejection of our extractions is 0.6-7.7% across tasks, far below error rates we measure on widely used curated counterparts (e.g., 16.5% on BBB_Martins, 7.3% on Bioavailability_Ma). Beyond scale and accuracy, the supporting passages carry nuance tabular databases discard – e.g., oral bioavailability may depend on fed vs. fasted state. Together, the corpus, retrieval, and agent establish a foundation for AI-driven therapeutic design. Code and datasets: https://github.com/starling-labs/starling.

阅读与讨论 → 访问原文 →

05.

arXiv (CS.CL) 2026-06-19 DOI: arXiv:2604.18105

NIM4-ASR: Towards Efficient, Robust, and Customizable Real-Time LLM-Based ASR

作者:

Yuan Xie↗Jiaqi Song↗Guang Qiu↗Xianliang Wang↗Kai Qiao↗Junfeng Yuan↗Shengqing Liu↗Yi Zhang↗Bowen Chen↗Ming Lei↗Jie Gao↗Jie Wu↗…

Integrating large language models (LLMs) into automatic speech recognition (ASR) has become a mainstream paradigm in recent years. Although existing LLM-based ASR models demonstrate impressive performance on public benchmarks, their training remains predominantly data-driven, leaving key practical challenges insufficiently addressed – particularly limited downward scalability in resource-constrained deployments and hallucinations under acoustically challenging conditions. To address these issues, we present NIM4-ASR, a production-oriented LLM-based ASR framework optimized for both efficiency and robustness. Grounded in a principled delineation of functional roles between the encoder and the LLM, we redesign the multi-stage training paradigm to align each module with its intended capability boundary. Specifically, we reformulate the pre-training architecture and objective to mitigate the modality gap and improve parameter efficiency; introduce an iterative asynchronous SFT stage to preserve acoustic fidelity and constrain representation drift; and design an ASR-specialized reinforcement learning stage to further enhance recognition quality and robustness. We additionally incorporate a suite of production-oriented optimizations, including robustness under noisy and silent conditions, real-time streaming inference, and hotword customization via retrieval-augmented generation (RAG). Experiments show that NIM4-ASR achieves state-of-the-art performance on multiple public benchmarks with merely 2.3B parameters, while substantially outperforming larger-scale competitors on internal benchmarks – particularly in entity-intensive real-world scenarios. NIM4-ASR further supports million-scale hotword customization via RAG with sub-millisecond retrieval latency, enabling efficient adaptation to emerging entities and personalized user requirements.

阅读与讨论 → 访问原文 →

06.

arXiv (CS.CL) 2026-06-15 DOI: arXiv:2505.04671

Reward-SQL: Boosting Text-to-SQL via Stepwise Execution-Aware Reasoning and Process-Supervised Rewards

作者:

Yuxin Zhang↗Meihao Fan↗Ju Fan↗Mingyang Yi↗Yuyu Luo↗Guoliang Li↗Bin Wu↗Wenchao Zhou↗

Recent advances in large language models (LLMs) trained with reinforcement learning (RL) have improved Text-to-SQL performance. However, RL-based approaches still struggle with complex queries due to two key limitations: insufficient stepwise execution-aware reasoning grounded in database feedback, and the lack of process-level rewards for guiding reasoning optimization. To address these issues, we propose CoCTE, a divide-and-conquer and execution-aware reasoning framework that progressively composes SQL queries through intermediate view validation and structured Common Table Expressions (CTEs), improving both accuracy and interpretability. To realize a CoCTE reasoning process, we develop Reward-SQL, a unified approach with three stages: (1) model initialization, which equips LLMs with structured CoCTE reasoning capabilities; (2) process reward design, which delivers fine-grained, execution-aware supervision; and (3) process-supervised RL and inference, which integrates process rewards into training and guides the inference stage by process rewards. This paper addresses the core challenges in Reward-SQL and makes the following contributions. We introduce a process reward model (PRM) that combines execution-aware trajectory scoring with entropy-based step weighting, providing dense and interpretable supervision across reasoning steps. We integrate PRM into both RL training and inference stages, stabilizing optimization and improving trajectory exploration with process-level signals. Experiments show that Reward-SQL significantly outperforms baselines with comparable model sizes, and exhibits strong cross-domain generalization.

阅读与讨论 → 访问原文 →

07.

arXiv (CS.CL) 2026-06-19 DOI: arXiv:2606.19348

DeepSeek-V4: Towards Highly Efficient Million-Token Context Intelligence

作者:

DeepSeek-AI↗Anyi Xu↗Bangcai Lin↗Bing Xue↗Bingxuan Wang↗Bingzheng Xu↗Bochao Wu↗Bowei Zhang↗Chaofan Lin↗Chen Dong↗Chenchen Ling↗Chengda Lu↗…

We present a preview version of DeepSeek-V4 series, including two strong Mixture-of-Experts (MoE) language models – DeepSeek-V4-Pro with 1.6T parameters (49B activated) and DeepSeek-V4-Flash with 284B parameters (13B activated) – both supporting a context length of one million tokens. DeepSeek-V4 series incorporate several key upgrades in architecture and optimization: (1) a hybrid attention architecture that combines Compressed Sparse Attention (CSA) and Heavily Compressed Attention (HCA) to improve long-context efficiency; (2) Manifold-Constrained Hyper-Connections (mHC) that enhance conventional residual connections; (3) and the Muon optimizer for faster convergence and greater training stability. We pre-train both models on more than 32T diverse and high-quality tokens, followed by a comprehensive post-training pipeline that unlocks and further enhances their capabilities. DeepSeek-V4-Pro-Max, the maximum reasoning effort mode of DeepSeek-V4-Pro, redefines the state-of-the-art for open models, outperforming its predecessors in core tasks. Meanwhile, DeepSeek-V4 series are highly efficient in long-context scenarios. In the one-million-token context setting, DeepSeek-V4-Pro requires only 27% of single-token inference FLOPs and 10% of KV cache compared with DeepSeek-V3.2. This enables us to routinely support one-million-token contexts, thereby making long-horizon tasks and further test-time scaling more feasible. The model checkpoints are available at https://huggingface.co/collections/deepseek-ai/deepseek-v4.

阅读与讨论 → 访问原文 →

08.

arXiv (CS.AI) 2026-06-17 DOI: arXiv:2606.18023

LoopCoder-v2: Only Loop Once for Efficient Test-Time Computation Scaling

作者:

Jian Yang↗Shawn Guo↗Wei Zhang↗Tianyu Zheng↗Yaxin Du↗Haau-Sing Li↗Jiajun Wu↗Yue Song↗Yan Xing↗Qingsong Cai↗Zelong Huang↗Chuan Hao↗…

arXiv:2606.18023v1 Announce Type: cross Abstract: Looped Transformers scale latent computation by repeatedly applying shared blocks, but sequential looping increases latency and KV-cache memory with the loop count. Parallel loop Transformers (PLT) alleviate this cost through cross-loop position offsets (CLP) and shared-KV gated sliding-window attention, making loop count a practical design choice. We therefore study PLT loop-count selection through a gain–cost view: an extra loop may refine representations, but CLP also introduces a positional mismatch at each loop boundary. We instantiate this study by training LoopCoder-v2, a family of 7B PLT coders with different loop counts, from scratch on 18T tokens, followed by matched instruction tuning and evaluation. Empirically, the two-loop variant delivers broad gains over the non-looped baseline across code generation, code reasoning, agentic software engineering, and tool-use benchmarks, improving SWE-bench Verified from 43.0 to 64.4 points and Multi-SWE from 14.0 to 31.0 points. In contrast, variants with three or more loops regress, revealing a strongly non-monotonic loop-count effect. Our diagnostics show that loop 2 provides the main productive refinement, while later loops yield diminishing, oscillatory updates and reduced representational diversity. Because the CLP-induced mismatch remains roughly fixed as refinement gains shrink, the offset cost increasingly dominates. This gain–cost trade-off explains PLT's saturation at two loops and provides diagnostics for loop-count selection.

阅读与讨论 → 访问原文 →

09.

arXiv (CS.CV) 2026-06-17 DOI: arXiv:2602.18746

Bridging Modality Disconnect in Self-Reflection via Closed-Loop Visually Grounded Verification

作者:

Haoyu Zhang↗Yuwei Wu↗Pengxiang Li↗Xintong Zhang↗Zhi Gao↗Rui Gao↗Mingyang Gao↗Che Sun↗Yunde Jia↗

In the era of Vision-Language Models (VLMs), enhancing multimodal reasoning capabilities remains a critical challenge, particularly in handling ambiguous or complex visual inputs, where initial inferences often lead to hallucinations or logic errors. Existing VLMs often produce plausible yet ungrounded answers, and even when prompted to "reflect", their corrections may remain detached from the image evidence. To address this, we propose the MIRROR framework for Multimodal Iterative Reasoning via Reflection On visual Regions. By embedding visual reflection as a core mechanism, MIRROR is formulated as a closed-loop process comprising draft, critique, region-based verification, and revision, which are repeated until the output is visually grounded. To facilitate training of this model, we construct **ReflectV**, a visual reflective dataset for multi-turn supervision that explicitly contains reflection triggers, region-based verification actions, and answer revision grounded in visual evidence. Experiments on both general vision-language benchmarks and representative vision-language reasoning benchmarks show that MIRROR improves correctness and reduces visual hallucinations, demonstrating the value of training reflection as an evidence-seeking, region-aware verification process rather than a purely textual revision step.

阅读与讨论 → 访问原文 →

10.

arXiv (CS.CV) 2026-06-16 DOI: arXiv:2606.15037

ReportQA: QA-Based Radiology Report Evaluation

作者:

Yiming Shi↗Shaoshuai Yang↗Xi Chen↗Haolin Li↗Hengyu Zhang↗Che Jiang↗Kaiwen Wang↗Xun Zhu↗Dong Xie↗Fei Wang↗Dejing Dou↗Miao Li↗…

Radiology report evaluation is essential for advancing automated report generation. Natural language generation metrics have limited clinical relevance. Clinical efficacy (CE) metrics evaluate important medical findings, but focus mainly on presence and cover only a limited set of entities. Due to heavy reliance on manual annotations, it is difficult for CE metrics to extend clinical entities or attributes. In clinical practice, radiology reports serve as a medium for information transfer. Clinicians use them to perform downstream diagnostic tasks without directly inspecting images. Based on this insight, we propose ReportQA, a clinical-related and flexible radiology report evaluation framework, supporting detailed quantitative analysis of radiology report generation systems. We first collect datasets covering multiple imaging modalities and anatomical regions. We then construct knowledge trees of clinical entities and attributes with radiologist guidance, and use large language models (LLMs) to extract structured information from raw reports. Next, we generate QA pairs from predefined templates and apply quality control through self-filtering and report-based filtering. During evaluation, the report is treated as context, and an LLM acts as a judge model to answer the QA pairs. Based on the resulting QA accuracy, we introduce QAScore metric. Compared with existing metrics, QAScore shows better alignment with radiologist judgments. Experiments on multiple state-of-the-art vision-language models reveal that current report-based inference paradigms struggle to learn fine-grained clinical representations and exhibit strong negative prior biases. In contrast, question-driven inference provides a more effective alternative. For reproducibility and extensibility, we release the knowledge trees, structured reports, and QA pairs, along with the pipeline code for QA construction and evaluation.

阅读与讨论 → 访问原文 →

11.

arXiv (CS.CV) 2026-06-15 DOI: arXiv:2606.14504

Scratched Lenses, Shifted Depth: Passive Camera-Side Optical Attacks

作者:

Qinlin He↗Zeming Zhuang↗Yongji Wu↗Lan Zhang↗Xiaoyong↗Yuan↗

Physical adversarial attacks on vision systems are typically studied through scene manipulation, such as adversarial patches or projections, where the adversary controls what the camera observes. Camera-side attacks using stickers or auxiliary optics have also been explored, but they treat attacks as image-space perturbations from designed patterns. This misses how physical imperfections interact with scene-dependent lighting and optics. We identify a threat: passive lens-side damage that is persistent yet trigger-conditioned, producing optical artifacts that bias geometric inference under particular visual conditions. We instantiate this threat through Scratch-induced Lens Adversarial Streak Hijacking SLASH, a physical-world attack caused by small scratches on a camera lens or protective cover. Scratches interact with bright light sources and specular reflections to create structured streak artifacts that distort depth cues. Since the perturbation is fixed in the optical path but triggered by the scene, it is both persistent and selective. We formulate the attack in optical space, model the scratch pattern as a trigger-conditioned optical channel, and optimize one fixed configuration across diverse viewing conditions. We evaluate SLASH on monocular depth estimation and monocular 3D object detection in digital and real-world settings. Under the fixed-scratch constraint, directional depth shifts reach up to 32% relative error for monocular depth estimation, with consistent effects on monocular 3D object detection. Physical experiments confirm transfer to real camera recordings, inducing depth shifts above the model's natural prediction baseline. These findings reveal an attack surface where benign-looking hardware imperfections act as latent, scene-triggered adversarial mechanisms, challenging assumptions about physical robustness and motivating defenses for secure vision systems.

阅读与讨论 → 访问原文 →

12.

arXiv (CS.AI) 2026-06-16 DOI: arXiv:2606.16070

Mind-Studio: Executable World Models with Lookahead Evaluation for Partially Observable Games

作者:

Yifei Dong↗Mingen Zheng↗Linquan Wu↗Jeff Z. Pan↗Jiaxin Bai↗

arXiv:2606.16070v1 Announce Type: new Abstract: World-model synthesis aims to turn interaction experience into an internal model of environment dynamics. Existing symbolic approaches often fit observed transitions or mixtures of local rules, but they do not produce a complete executable program that can run independently of the real environment. We present Mind-Studio, a framework that synthesizes executable pygame-style world models from state-action-next-state trajectories using large language models. Mind-Studio combines entropy-selected traces with a lightweight game skill file containing object, action, and static scene information extracted from screenshots. We evaluate synthesis quality with a K-step lookahead fidelity protocol that compares generated world-model rollouts against Real-ALE rollouts from the same state. On Montezuma's Revenge, Mind-Studio improves chosen-action next-state prediction from 0.3% for PoE-World to 48.7% while verifying 5 of 8 subgoals; across Alien, Assault, and Skiing, it achieves stronger branch-level fidelity than prior learned lookahead sources.

阅读与讨论 → 访问原文 →

13.

arXiv (CS.AI) 2026-06-17 DOI: arXiv:2602.07429

Brep2Shape: Boundary and Shape Representation Alignment via Self-Supervised Transformers

作者:

Yuanxu Sun↗Yuezhou Ma↗Haixu Wu↗Guanyang Zeng↗Muye Chen↗Jianmin Wang↗Mingsheng Long↗

arXiv:2602.07429v2 Announce Type: replace-cross Abstract: Boundary representation (B-rep) is the industry standard for computer-aided design (CAD). While deep learning shows promise in processing B-rep models, existing methods suffer from a representation gap: continuous approaches offer analytical precision but are visually abstract, whereas discrete methods provide intuitive clarity at the expense of geometric precision. To bridge this gap, we introduce Brep2Shape, a novel self-supervised pre-training method designed to align abstract boundary representations with intuitive shape representations. Our method employs a geometry-aware task where the model learns to predict dense spatial points from parametric Bézier control points, enabling the network to better understand physical manifolds derived from abstract coefficients. To enhance this alignment, we propose a Dual Transformer backbone with parallel streams that independently encode surface and curve tokens to capture their distinct geometric properties. Moreover, the topology attention is integrated to model the interdependencies between surfaces and curves, thereby maintaining topological consistency. Experimental results demonstrate that Brep2Shape offers significant scalability, achieving state-of-the-art accuracy and faster convergence across various downstream tasks.Code is available at this repository: https://github.com/thuml/Brep2Shape.

阅读与讨论 → 访问原文 →

14.

arXiv (CS.CV) 2026-06-16 DOI: arXiv:2604.28185

Visual Generation in the New Era: An Evolution from Atomic Mapping to Agentic World Modeling

作者:

Keming Wu↗Zuhao Yang↗Kaichen Zhang↗Shizun Wang↗Haowei Zhu↗Sicong Leng↗Zhongyu Yang↗Qijie Wang↗Sudong Wang↗Ziting Wang↗Zili Wang↗Hui Zhang↗…

Recent visual generation models have made major progress in photorealism, typography, instruction following, and interactive editing, yet they still struggle with spatial reasoning, persistent state, long-horizon consistency, and causal understanding. We argue that the field should move beyond appearance synthesis toward intelligent visual generation: plausible visuals grounded in structure, dynamics, domain knowledge, and causal relations. To frame this shift, we introduce a five-level taxonomy: Atomic Generation, Conditional Generation, In-Context Generation, Agentic Generation, and World-Modeling Generation, progressing from passive renderers to interactive, agentic, world-aware generators. We analyze key technical drivers, including flow matching, unified understanding-and-generation models, improved visual representations, post-training, reward modeling, data curation, synthetic data distillation, and sampling acceleration. We further show that current evaluations often overestimate progress by emphasizing perceptual quality while missing structural, temporal, and causal failures. By combining benchmark review, in-the-wild stress tests, and expert-constrained case studies, this roadmap offers a capability-centered lens for understanding, evaluating, and advancing the next generation of intelligent visual generation systems.

阅读与讨论 → 访问原文 →

15.

arXiv (CS.CL) 2026-06-16 DOI: arXiv:2605.21850

ACC: Compiling Agent Trajectories for Long-Context Training

作者:

Qisheng Su↗Zhen Fang↗Shiting Huang↗Yu Zeng↗Yiming Zhao↗Kou Shi↗Ziao Zhang↗Lin Chen↗Zehui Chen↗Lijun Wu↗Feng Zhao↗

Recent development of agents has renewed demand for long-context reasoning capacity of LLMs. However, training LLMs for this capacity requires costly long-document curation or heuristic context synthesis. We observe that agents produce massive trajectories when solving problems, invoking tools and receiving environment observations across many turns. The evidence needed to answer the original question is thus scattered throughout these turns, requiring integration of distant context segments. Nevertheless, standard agent SFT masks tool responses and only trains turn-level tool selection, creating a supervision blind spot where these scattered signals go unused. We propose Agent Context Compilation (ACC), which converts trajectories from search, software engineering, and database querying agents into long-context QA pairs that combine the original question with tool responses and environment observations gathered across multiple turns, training the model to answer directly without tool use. This makes the dependencies between the question and the evidence explicit, enabling direct supervision of long-context reasoning over distant segments without additional annotation. ACC is a simple but effective approach that can be combined with any existing long-context extension or training method, providing scalable supervised fine-tuning data. We validate ACC on long-range dependency modeling tasks through MRCR and GraphWalks, challenging benchmarks requiring cross-turn coreference resolution and graph traversal over extended contexts. Training Qwen3-30B-A3B with ACC achieves 68.3 on MRCR (+18.1) and 77.5 on GraphWalks (+7.6), results comparable to Qwen3-235B-A22B, while preserving general capabilities on GPQA, MMLU-Pro, AIME, and IFEval. Further mechanism analysis reveals that the ACC-trained model exhibits task-adaptive attention restructuring and expert specialization.

阅读与讨论 → 访问原文 →

16.

arXiv (CS.CL) 2026-06-19 DOI: arXiv:2606.19659

SAGE-OPD: Selective Agent-Guided Intervention for Multi-Turn On-Policy Distillation

作者:

Yuhang Zhou↗Lizhu Zhang↗Yifan Wu↗Mingyi Wang↗Bo Peng↗Jiayi Liu↗Xiangjun Fan↗Zhuokai Zhao↗

On-policy distillation (OPD) improves student models by training them on trajectories induced by their own policy, making it a promising approach for mitigating exposure bias in agent training. However, most OPD studies focus on single-turn settings, while realistic LLM agents interact with environments over multiple turns. In this regime, early errors can alter future observations and compound across the trajectory, and standard dense token-level OPD becomes brittle, as it may over-penalize semantically valid alternatives, reinforce local degeneracies such as repeated actions, and propagate unreliable teacher supervision on off-distribution histories. We propose SAGE-OPD, a verifier-free selective intervention framework specifically designed for multi-turn OPD. Instead of applying teacher supervision uniformly across all turns, SAGE-OPD first observes environment feedback and uses teacher judgment to decide whether each student response should be skipped or intervened on. To further address compounding errors, SAGE-OPD weights token-level distillation by teacher confidence, reducing the influence of uncertain teacher distributions on corrupted or ambiguous histories. Finally, SAGE-OPD applies loss normalization to preserve the overall loss scale of standard OPD while retaining selective turn-level weighting. Experiments on agent tasks show that SAGE-OPD consistently improves over baselines, achieving up to a 13.3% relative improvement in ALFWorld unseen success rate over standard OPD. Ablation studies further demonstrate that turn-level intervention, teacher confidence weighting, and loss normalization provide complementary benefits. Our results suggest that effective multi-turn OPD should remain on-policy, but teacher supervision should be selectively allocated to turns where intervention is necessary and reliable.

阅读与讨论 → 访问原文 →

17.

arXiv (CS.CL) 2026-06-12 DOI: arXiv:2606.05405

Agents' Last Exam

作者:

Yiyou Sun↗Xinyang Han↗Weichen Zhang↗Yuanbo Pang↗Tianyu Wang↗Yuhan Cao↗Yixiao Huang↗Chris Duroiu↗Haoyun Zhang↗Jeffrey Lin↗Weishu Zhang↗Tyler Zeng↗…

Recent AI systems have achieved strong results on a wide range of benchmarks, yet these gains have not translated into economically meaningful deployment across many professional domains. We argue that this gap is largely an evaluation problem: widely used benchmarks lack sustained performance measurement on real and economically valuable workflows. This paper introduces Agents' Last Exam (ALE), a benchmark designed to evaluate AI agents on long horizon, economically valuable, real world tasks with verifiable outcomes. Developed in collaboration with 250+ industry experts, ALE covers non-physical industries defined with reference to O*NET / SOC 2018 (the U.S. federal occupational taxonomy). It is organized around a task taxonomy with 55 sub fields grouped into 13 industry clusters covering 1K+ tasks. Current results show that the hardest tier remains far from saturated: across mainstream harness and backbone configurations, the average full pass rate is below 1%. ALE is designed as a living benchmark: its task pool grows continuously as new workflows and industries are onboarded. More broadly, ALE is intended not merely as another leaderboard, but as an instrument for closing the gap between benchmark success and GDP relevant impact.

阅读与讨论 → 访问原文 →

18.

arXiv (CS.CV) 2026-06-12 DOI: arXiv:2606.13674

RepWAM: World Action Modeling with Representation Visual-Action Tokenizers

作者:

Junke Wang↗Qihang Zhang↗Shuai Yang↗Yiming Luo↗Yujun Shen↗Zuxuan Wu↗Yu-Gang Jiang↗Yinghao Xu↗

This work presents RepWAM, a representation-centric world action model (WAM) built on representation visual-action tokenizers. Existing WAMs typically inherit reconstruction-oriented video tokenizers from pretrained video generation models. Although these tokenizers preserve visual fidelity, pixel reconstruction alone provides limited guidance for learning instruction-following dynamics that connect future prediction with robot control. To address this, we explore a semantic visual-action latent space for representation-centric world action modeling. Specifically, we train a representation visual-action tokenizer that maps visual inputs into aligned visual and latent action tokens. We then pretrain our WAM to jointly model future visual states and the latent actions that connect them under language instructions, followed by adaptation to real robot trajectories for closed-loop manipulation. Experiments on real-world manipulation tasks and simulation benchmarks show that RepWAM delivers strong performance across diverse manipulation settings, while ablations highlight the value of semantic visual-action tokenization over reconstruction-oriented alternatives. These results establish representation visual-action tokenization as a promising foundation for world action models and a step toward generalist robot policies. Code and weights will be available at https://github.com/wdrink/RepWAM.

阅读与讨论 → 访问原文 →

19.

arXiv (CS.CL) 2026-06-15 DOI: arXiv:2606.14199

OdysSim: Building Foundation Models for Human Behavior Simulation

作者:

Xuhui Zhou↗Weiwei Sun↗Weihua Du↗Jiarui Liu↗Haojia Sun↗Qianou Ma↗Tongshuang Wu↗Yiming Yang↗Maarten Sap↗

Large language models are increasingly deployed as human simulators for interactive evaluation and social simulation. Yet helpfulness-driven post-training pulls them toward a homogeneous, overly agreeable assistant register, creating a behavioral Sim2Real gap. We present OdysSim, the largest open systematic investigation of behavioral foundation models, i.e., models trained to simulate human behavior at scale. We propose SOUL, a taxonomy of five capability axes (CONV, SS, COG, ROLE, EVAL) that unifies 62 datasets and 23 benchmark tasks under one framework. Specifically, we curate the OdysSim corpus (21.4M interactions, 10B tokens, retrofitted with back-generated social contexts), construct the SOUL-Index benchmark, and develop an end-to-end training recipe combining midtraining, task-specific RL, and expert distillation. The resulting open 8B OSim model ranks first or tied-first on 8 of 23 tasks, outperforming any individual frontier model by this count, with the strongest gains on conversational and social tasks. Its outputs are also more human-like in length, formatting, and word choice, and it transfers zero-shot to out-of-distribution user simulation on $\tau$-bench, nearly matching real users on reaction alignment (93.2 vs. 93.5). We further show that LLM-as-judge RL induces reward-hacking patterns, and that our detectors can mitigate them during post-training. Together, our findings suggest that behavioral foundation models require rethinking the LLM training paradigm. We release all artifacts to support future research.

阅读与讨论 → 访问原文 →

20.

arXiv (CS.CL) 2026-06-15 DOI: arXiv:2606.07040

Beyond Rubrics: Exploration-Guided Evaluation Skills for Reward Modeling

作者:

Xing Yue↗Linjuan Wu↗Daoxin Zhang↗Yongliang Shen↗Weiming Lu↗

Open-ended reward modeling requires judges that can follow subtle, domain-specific preferences when verifiable answers are unavailable. Existing rubric-based methods often address this by generating criteria online for each query, but the extra generation step can add inference overhead and produce rigid or misaligned guidance. We introduce Eval-Skill, an exploration-guided method that synthesizes reusable evaluation skills for reward modeling and reframes reward guidance as context evolution rather than parameter training or per-query rubric generation. Using only 100 cases per domain for skill evolution, Eval-Skill synthesizes reusable domain-level evaluation skills through two progressive stages, workflow generation followed by principle generation, with exploration and selection interleaved across both stages. Once generated, a skill is directly injected into the judge context. Across multiple RM benchmarks, Eval-Skill consistently improves diverse judge backbones; on RewardBench 2, it yields significant gains over vanilla judging for each main backbone (+13.44% for Qwen3-8B, and 18.51% for DeepSeek-V4-Flash). Further analyses of evolution-time scaling, generalizability, and transferability show that compact evaluation skills offer an efficient new paradigm for LLM-based evaluation. Code is available at https://github.com/xing-stellus-yue/Eval-Skill.

阅读与讨论 → 访问原文 →

21.

arXiv (CS.AI) 2026-06-16 DOI: arXiv:2606.15306

LatentGym: A Testbed For Cross-Task Experiential Learning With Controllable Latent Structure

作者:

Daksh Mittal↗Tommaso Castellani↗Thomson Yen↗Naimeng Ye↗Fangyu Wu↗Minghui Chen↗Tiffany Cai↗Emmanouil Koukoumidis↗William Zeng↗Hongseok Namkoong↗

arXiv:2606.15306v1 Announce Type: cross Abstract: We envision continually learning agentic systems that become more useful over time: as they encounter sequences of related tasks, they should infer the hidden structure shared across those tasks and use it to improve future decisions. This cross-task experiential learning capability is pivotal in domains such as personalization and interactive assistance, but existing training/evaluation frameworks do not provide shared, controllable latent structures and cannot measure whether or why agents improve. We introduce LatentGym: a controllable suite in which each environment is organized around a ground-truth latent variable governing the structure across tasks. Our construction yields metrics that separate exploration (whether the agent's actions gather information about the latent) from exploitation (whether the agent uses what it has gathered). We demonstrate our suite on empirical studies addressing three questions: how and why frontier models fail to adapt across related tasks; whether post-training on related task sequences improves general cross-task adaptation, and where those gains come from; and how design choices such as inter-task feedback shape training dynamics and generalization. Together, these results establish a controlled foundation for studying how LLM agents learn from experience across tasks, and for designing agents that adapt more reliably in sequential, personalized, and interactive settings.

阅读与讨论 → 访问原文 →

22.

arXiv (CS.AI) 2026-06-15 DOI: arXiv:2606.14409

Hy-Embodied-0.5-VLA: From Vision-Language-Action Models to a Real-World Robot Learning Stack

作者:

He Zhang↗Lingzhu Xiang↗Haitao Lin↗Zeyu Huang↗Minghui Wang↗Dingyan Zhong↗Yubo Dong↗Yihao Wu↗Yongming Rao↗Dongsheng Zhang↗Wanjia He↗Ling Chen↗…

arXiv:2606.14409v1 Announce Type: cross Abstract: In this report, we present Hy-Embodied-0.5-VLA, abbreviated as HyVLA-0.5, an end-to-end system that spans the full robot learning stack: data collection, model design, continued pre-training and supervised fine-tuning, RL post-training, and real-world deployment. Each component serves a distinct role in this stack.

阅读与讨论 → 访问原文 →

23.

arXiv (CS.LG) 2026-06-19 DOI: arXiv:2606.04101

UltraEP: Unleash MoE Training and Inference on Rack-Scale Nodes with Near-Optimal Load Balancing

作者:

Xinming Wei↗Chao Jin↗Tuo Dai↗Yinmin Zhong↗Shan Yu↗Chengxu Yang↗Bingyang Wu↗Zili Zhang↗Jing Mai↗Qianchao Zhu↗Zhouyang Li↗Yuliang Liu↗…

arXiv:2606.04101v3 Announce Type: replace-cross Abstract: Large-scale expert parallelism (EP) is becoming pivotal for training and serving frontier MoE models, but it also amplifies device-level expert load imbalance into compute stragglers, token all-to-all bottlenecks, and activation-memory spikes. Existing balancers redistribute experts periodically based on historical load, which becomes unreliable for production deployments with non-stationary load patterns. We present UltraEP, the first exact-load, real-time balancer for large-EP MoE training and serving prefill on rack-scale nodes (RSNs). Leveraging the extended scale-up connectivity among dozens of GPUs within RSNs, UltraEP rebalances every microbatch and layer on critical paths, which requires nontrivial co-design of plan solving and expert replication communication to minimize exposed overhead. To this end, UltraEP eagerly reacts to post-gating load with an efficient quota-driven planner, and executes the resulting irregular expert-state transfers with RSN-native persistent tile streaming and relay-based fan-out mitigation. We evaluate UltraEP in a multi-RSN deployment of up to 256 GPUs, using cutting-edge MoE models from 106B to 671B parameters. Averaged across training and serving, UltraEP achieves 94.3% of the force-balanced ideal throughput, delivering 1.49$\times$ improvement over no-balancing, while reducing the final inter-rank imbalance from 1.30$-$4.01 to 1.01$-$1.04.

阅读与讨论 → 访问原文 →

24.

arXiv (CS.CL) 2026-06-11 DOI: arXiv:2408.02600

BioMamba: Domain-Adaptive Biomedical Language Models

作者:

Ling Yue↗Mingzhi Zhu↗Sixue Xing↗Yunning Cao↗Yanbo Wang↗Shimin Shan↗Jinfei Liu↗Vijil Chenthamarakshan↗Shaowu Pan↗Payel Das↗Tianfan Fu↗

Background. Biomedical language models should improve performance on biomedical text while retaining general-language-modeling fluency. For Mamba-based models, this trade-off has not been systematically studied across biomedical literature and clinical text. Methods. We developed BioMamba, a family of biomedical Mamba2 models at five scales obtained by continued pretraining of released public Mamba2 checkpoints on a balanced 80%/10%/10% mixture of PubMed abstracts, the Colossal Clean Crawled Corpus (C4), and Wikipedia. The contribution is the adaptation recipe and the accompanying open-weight checkpoints. Results. Across five scales, BioMamba consistently lowered PubMed perplexity, improved Wikipedia-style held-out perplexity by 1.46-4.72 PPL, and left C4 perplexity essentially unchanged. On six out-of-domain multiple-choice benchmarks, BioMamba stayed within +/-3 percentage points of Mamba2 with no systematic regression. After supervised fine-tuning, BioMamba+SFT matched or exceeded Mamba2+SFT on MIMIC-IV note completion and discharge summary generation at every evaluated scale, and improved PubMedQA at every scale. The strongest model (BioMamba-2.7B) reached a PubMed perplexity of 5.28 and accuracies of 90.24% and 73.00% on BioASQ and PubMedQA, respectively. Conclusions. A balanced domain-adaptive continued pretraining recipe strengthens Mamba2 language models on biomedical literature and clinical text while preserving general-language-modeling fluency.

阅读与讨论 → 访问原文 →

25.

arXiv (CS.LG) 2026-06-17 DOI: arXiv:2606.18105

OmniPlan: An Adaptive Framework for Timely and Near-Optimal Network Planning Optimization

作者:

Longlong Zhu↗Jiashuo Yu↗Zedi Chen↗Yuhan Wu↗Zhifan Jiang↗Yuchen Xian↗Yimeng Liu↗Jiajie Su↗Shaopeng Zhou↗Xingyuan Li↗Hongyan Liu↗Xuan Liu↗…

arXiv:2606.18105v1 Announce Type: cross Abstract: Network planning optimization is a fundamental problem across diverse domains, including transportation systems, communication networks, and power grids. It requires simultaneous optimization of multiple competing objectives under complex constraints. Existing network planning optimization frameworks rely on mixed integer programming (MIP) solvers, heuristics, and deep reinforcement learning (DRL) models to compute planning decisions. However, they lack effective adaptability to diverse and dynamic user intents, thus leading to the trade-off between execution time and optimality. In this paper, we propose OmniPlan, an adaptive framework that achieves both timeliness and near-optimality in network planning optimization. To achieve the adaptability lacking in existing solutions, OmniPlan employs a large language model (LLM)-based interpreter to convert heterogeneous natural-language intents into a unified and quantifiable user-preference vector. Then it employs a mixture-of-experts architecture that integrates MIP solvers, heuristics, and DRL models as specialized experts, where OmniPlan adapts to diverse intents by dynamically selecting timely and near-optimal experts. Finally, it incorporates a DRL-based expert configuration module that fine-tunes optimization objective weights to align planning decisions with user-specific preferences. We evaluate OmniPlan with a representative real-world workload, i.e., distributed machine learning (ML), where we leverage OmniPlan to offload a wide spectrum of ML inference tasks, e.g., decision trees, SVM, naive Bayes, XGBoost, and random forests, onto a network of hardware devices. Our experiments on a real-world testbed indicate that OmniPlan achieves near-optimal and low-execution-time offloading for real-world ML inference tasks, reducing latency by up to 97.8\% and network device resource consumption by up to 11.5\%.

阅读与讨论 → 访问原文 →