论文

结果表明，由 LLM 编排的多 Agent 系统可将传统 AutoML 扩展为可信、自适应且面向生产的 BDaaS 生命周期自动化。The results suggest that LLM-orchestrated multi-agent systems can extend conventional AutoML toward trustworthy, adaptive, and production-oriented BDaaS lifecycle automation.

提出一种针对 LLM Agent 的自动化提示优化框架，将"观测到动作"流水线分解为目标条件描述子 Agent 与动作选择 Agent，并通过由 LLM 驱动、以环境回报为指导的进化循环迭代优化各模块提示。An automated prompt optimization framework for LLM agents that decomposes the observation-to-action pipeline into a goal-conditioned descriptor agent and an action selection agent, and iteratively refines each module's prompt through an LLM-driven evolutionary loop guided by environment returns is introduced.

本文提出 MedEasy，一个多 Agent 系统，通过患者对话、临床操作、决策提交、文档记录与反馈来组织虚拟患者练习，为使用案例特定标准连接情境化实践的 AI 辅助职业训练系统贡献了设计启示。MedEasy is presented, a multi-agent system that organizes virtual-patient practice through patient dialogue, clinical actions, decision submission, documentation, and feedback that contributes design implications for AI-supported professional training systems that use case-specific standards to connect situated practice.

本文提出一个面向规则驱动查询改写的自进化框架，无需任何参数训练即可增强 BM25，并揭示 LLM 利用先前实验结果的能力以及其对规则消除的内在知识，在通过自进化精炼规则集方面起到关键作用。This work proposes a self-evolving framework for rule-driven query rewriting that enhances BM25 without any parameter training, and reveals that LLM's capabilities to leverage previous experimental results and its intrinsic knowledge of rule elimination play critical roles in refining the rule set via self-evolution.

RL-Index 被提出，是一个将检索索引推理建模为强化学习问题的索引框架，能持续提升检索与下游问答性能，同时显著降低在线推理延迟。RL-Index is proposed, an indexing framework that formulates retrieval index reasoning as a reinforcement learning problem that consistently improves both retrieval and downstream question-answering performance, while significantly reducing online inference latency.

本文认为，Agentic AI 的近期价值不在于完全自主或削减人力，而在于面向简单与中等复杂度业务流程的受控部分自主。It is argued that the near term value of Agentic AI does not lie in full autonomy or workforce reduction, but in controlled partial autonomy for simple and medium complexity business processes.

本文引入 trace-economic underwriting，将工具调用 trace 映射为客户风险敞口与可索赔损失，并以此表示用于定价、控制与风险转移，使用确定性经济标签而非 LLM 评判器。T trace-economic underwriting is introduced, which maps tool-use traces to customer exposure and claimable loss, then uses this representation for pricing, control, and risk transfer, and uses deterministic economic labels rather than an LLM judge.

新加坡 AI Safety Institute 与韩国 AI Safety Institute 联合评估了涵盖客服、DevOps、网页自动化以及企业与个人生产力场景下 12 项真实非对抗任务中的 Agent 数据泄露问题，表明操作性数据泄露是与对抗性数据外泄不同的一阶 Agent 安全问题。A joint evaluation by the Singapore AI Safety Institute and the Korea AI Safety Institute examining agent data leakage in 12 realistic, non-adversarial tasks spanning customer support, DevOps, web automation, and enterprise and personal productivity indicates that operational data leakage is a first-order agent-safety concern distinct from adversarial exfiltration.

对 Agent 行为的分析揭示了长视野经济交互中的显著差异：表现更好的模型与其他企业的沟通更为活跃，而 Claude Haiku 4.5 则表现出 idle-drift 失效模式，在生成连贯评估与规划的同时仍反复选择不行动。Analysis of agent behavior reveals substantial differences in long-horizon economic interaction: higher-performing models communicate more actively with other firms, whereas Claude~Haiku~4.5 exhibits an idle-drift failure mode, repeatedly choosing inaction despite producing coherent assessments and plans.

本文提出 MAGE-RAG，一个面向长文档多模态问答的多粒度自适应图证据框架，并建立了涵盖 Direct MLLM、Text RAG、Page-level Visual RAG 与 Graph/Agentic RAG 的统一比较与分析协议。This paper proposes MAGE-RAG, a multigranular adaptive graph evidence framework for long-document multimodal QA, and establishes a unified comparison and analysis protocol covering Direct MLLM, Text RAG, Page-level Visual RAG, and Graph/Agentic RAG.

本文理论上证明，归一化离散 Ricci Flow 可通过识别边权中的不同渐近行为来检测社区结构，并支持移除相对于 query 节点具有大权重与负 Ricci 曲率的"噪声"文档片段。It is theoretically prove that normalized discrete Ricci flow can detect community structures by identifying distinct asymptotic behaviors in edge weights, and supports the removal of ``noisy''document chunks characterized by large weights and negative Ricci curvature relative to the query node.

在 MS MARCO 与真实生产流量上的结果表明，自适应检索 cardinality 能够在不降低检索质量的前提下提升检索效率。Results on both MS MARCO and real-world production traffic suggest that adaptive retrieval cardinality can improve retrieval efficiency without degrading retrieval quality.

CoRe（Context Relevance）被提出。该系统在一个大型短视频搜索引擎中每周重新部署，持续运行超过五个月，使用已部署的多模态相关性模型作为源，并以镜像生产融合代数的乘性比值形式来缩小仿真与生产之间的差距。CoRe (Context Relevance) is presented, such a system, redeployed weekly for over five months in a major short-video search engine, using the deployed multimodal relevance model as its source and a multiplicative ratio form mirroring the production fusion algebra to close the simulation-production gap.

CQC-RAG 框架被提出，它协同设计查询级多样性注入与跨查询一致性评估，无需外部监督即可实现自我评估，验证了跨查询一致性在过滤噪声引发幻觉方面的有效性。CQC-RAG, a framework that co-designs query-level diversity injection with cross-query consistency evaluation and enables self-evaluation without external supervision, is introduced, validating the effectiveness of cross-query consistency for filtering noise-induced hallucinations.

AgentOdyssey 被提出，这是一个新颖的评估框架，通过程序化方式生成包含丰富实体、世界动态和长视野任务的开放式文本游戏，并发现短期记忆对多种智能体范式均有益，是智能体测试时训练的重要组成部分。AgentOdyssey is introduced, a novel evaluation framework that procedurally generates open-ended text games with rich entities, world dynamics, and long-horizon tasks and finds that short-term memory benefits multiple agent paradigms and is an important component of agent test-time training.

结果表明，交易历史中编码的行为知识可以被语言模型有效学习，为客户数字孪生和行为模拟提供了可扩展的基础。The results demonstrate that behavioral knowledge encoded in transaction histories can be effectively learned by language models, providing a scalable foundation for customer digital twins and behavior simulation.

FRAMe 展示了先进 LLM 如何被部署用于以人为本的任务规划，将自然语言指令转化为安全、高效且灵活的飞行路线。FRAMe signifies how advanced LLMs can be deployed for human-centric mission planning, translating natural language instructions into safe, efficient, and flexible flight routes.

本文提出了一种动作分级伤害评估量表，按七级有序量表对智能体的工具调用轨迹进行打分，评判依据包括所执行动作的可逆性、是否越界涉及其他方以及是否扩大了权限。An action-graded harm rubric is introduced that scores an agent's tool-call trajectory on a seven-level ordinal scale according to whether the executed action was reversible, whether it crossed scope to reach another party, and whether it expanded privilege.

本文在视觉语言导航、具身问答和语言条件操控任务上评估了三种 AAS 变体，覆盖四个具身执行器，结果表明架构级搜索能在具身任务上产生可部署且具有方向性的成功率提升，而其中一个看似得分较高的候选因存在泄漏而被判定为无效。This work evaluates three AAS variants across four embodied executors spanning vision-language navigation, embodied question answering, and language-conditioned manipulation, and shows that architecture-level search can produce deployable and directional success-rate gains on embodied tasks, while one apparent high-scoring candidate is rejected as leak-bearing.

提出 Single-rollout Asynchronous Optimization（SAO），用于解决异步 RL 中的稳定性与 off-policy 难题，可稳定训练一千步，并在 Agentic 编码与推理基准上一致优于 GRPO 及其变体。Single-rollout Asynchronous Optimization (SAO) is presented to address the stability and off-policy challenges in asynchronous RL and is able to train stably for one thousand steps and consistently outperform GRPO and its variants on agentic coding and reasoning benchmarks.

提出 Sparse Delta Memory，一种通过稀疏寻址方案将门控线性 RNN 隐状态容量扩展数个数量级的架构，在上下文学习与长上下文检索任务上显著提升性能。Sparse Delta Memory is introduced, an architecture that scales the hidden state of gated linear RNNs to orders of magnitude higher capacity using a sparse addressing scheme and significantly improves performance on in-context learning and long-context retrieval tasks.

提出 AgentLens，一个面向交互式代码 Agent 的生产级评估基准，将形式化验证（在存在客观检查时）与 LLM 编写的轨迹评审及并排比较相结合，使每次运行都能给出关于分数为何如此的可读解释。This work presents AgentLens, a production-assessed benchmark for interactive code agents that pairs formal verification, where an objective check exists, with LLM-written trajectory reviews and side-by-side comparisons, so that each run yields a readable explanation of why the score is what it is.

提出 RoboTALES，一个学习任务对齐模拟未来并据此训练机器人策略的单阶段框架，引入两项关键创新：基于 LLM 的分层规划器，将复杂任务拆解为子目标序列以引导模型的"想象"；基于 VLM 的评判器，用于评估这些"想象"出的未来。This work proposes RoboTALES, a single-stage framework that learns task-aligned simulated futures and uses them to train robot policies and introduces two key innovations: a hierarchical LLM-based planner that breaks complex tasks into a sequence of subgoals to guide the model's imagination and a VLM-based critic that evaluates these ``imagined'' futures.

一种以 Token 为中心的双视图学习框架，在冻结的视觉 Transformer 主干中统一基于 prompt 的适配与跨视图融合，相比线性探测、仅 prompt 适配以及传统融合基线均取得稳定提升。A token-centric dual-view learning framework that unifies prompt-based adaptation and cross-view fusion within a frozen vision transformer backbone and demonstrates consistent improvements over linear probing, prompt-only adaptation, and conventional fusion baselines.

提出 OmniTacTune，一种策略无关的真实世界 RL 流程，通过残差修正将触觉反馈适配到预训练视觉策略，并在多种接触丰富任务、视觉基础策略与触觉表征间实现泛化。OmniTacTune is introduced, a policy-agnostic real-world RL pipeline that adapts tactile feedback to pretrained visual policies through residual correction and generalizes across diverse contact-rich tasks, visual base policies, and tactile representations.

给出一个具体且有实证支撑的像素级 EO 基础模型扩展方案：训练大型 encoder，按下游性能筛选，再蒸馏为灵活的学生模型。A concrete, empirically grounded recipe for scaling pixel-wise EO foundation models: train large encoders, select by downstream performance, and distil into flexible student models is given.

[摘要] Splash 是一个面向 MLLMs 的掩码隔离触觉对齐学习框架，它量化每个预训练参数的重要性，并将参数空间划分为休眠子空间与关键子空间，从而有效防止灾难性遗忘，确保非破坏性的模态扩展。Splash is presented, a mask-isolated tactile alignment learning framework for MLLMs that quantifies the significance of each pretrained parameter, and partitions the parameter space into a dormant and critical subspace, which effectively prevents catastrophic forgetting and ensures non-destructive modality expansion.

提出首个面向动态真实世界场景评估主动 agent 的能力驱动基准，并在模型与框架层面的全面比较表明，基模型能力与 agent 框架设计共同决定了真实世界环境中的性能表现。The first capability-driven benchmark designed to evaluate proactive agents in dynamic, real-world settings is introduced, and comprehensive comparisons across both models and frameworks show how base model capabilities and agent framework designs jointly shape performance in real-world environments.

提出 TokenWall，一种作用于 agent token 流的语义防火墙式运行时防御框架，证明语义运行时约束可在持久化 AI agents 上实现实用的安全性与效用性权衡。TokenWall is proposed, a runtime defense framework that acts as a semantic firewall over agent token flows, demonstrating that semantic runtime containment can achieve a practical security-utility trade-off for persistent AI agents.

提出 Contextuality（上下文性）——即 AI 系统自主访问用户累积知识资本的程度——作为 AI 介导不平等的一个维度，补充但不可化约为 Sharp 等人的框架。Contextuality -- the degree to which an AI system autonomously accesses a user's accumulated knowledge capital -- is proposed as a dimension of AI-mediated inequality that complements, but is not reducible to, the Sharp et al. framework.

本文提出一种面向历史数字图书馆管理的文档分析系统，支持即时知识建模，并促进生成更丰富、更全面的信息。This article presents a document analysis system designed for the management of historical digital libraries that supports on-the-fly knowledge modeling and facilitates the generation of richer and more comprehensive information.

PolyUQuest 是一个基于异构图构建的可验证、结构感知 Web RAG 框架，统一了页面间超链接拓扑、页面内 DOM 层级以及跨页面实体-关系知识，在答案正确性、覆盖度和忠实度上优于现有 RAG 系统，且每次查询消耗的 LLM tokens 显著更少。PolyUQuest, a verifiable, structure-aware web RAG framework built on a heterogeneous graph that unifies hyperlink topology between pages, DOM hierarchy within pages, and entity-relation knowledge across pages, outperforms existing RAG systems in answer correctness, coverage, and faithfulness, while consuming significantly fewer LLM tokens per query.

CineMobile 采用三重优化策略，通过蒸馏引导的剪枝方法得到一个紧凑而高效的模型，保留实现电影级效果所需的核心视频生成能力，证明了其在移动端图像到视频创作中的实用可行性。CineMobile adopts a three-fold optimization strategy, leveraging a distillation-guided pruning approach to derive a compact yet efficient model that retains the essential video generation capabilities required for cinematic effects, demonstrating its practical applicability for mobile-based image-to-video creation.

该审计揭示现有基准样本中 55% 可在无视觉输入或时序上下文的情况下被解决，并提出 Video-Oasis，一个用于系统性审计现有视频理解基准的可持续诊断套件。This audit reveals that 55\% of existing benchmark samples are solvable without visual input or temporal context, and introduces Video-Oasis, a sustainable diagnostic suite for systematically auditing existing video understanding benchmarks.

本文论证了稀疏组合监督与动宾学习的不对称性会助长物体驱动的捷径学习，并指出减少捷径诊断可提升组合泛化能力。This work argues that sparse compositional supervision and verb-object learning asymmetry can promote object-driven shortcut learning and reduces shortcut diagnostics and consequently improves compositional generalization.

本文提出 LongE2V，一种利用预训练视频扩散先验来联合处理基于事件视频重建、预测与帧间插值的新方法，并引入自回归展开与自适应上下文切换机制，以缓解超长序列中的时序漂移问题。This work proposes LongE2V, a novel approach that leverages pre-trained video diffusion priors to jointly handle event-based video reconstruction, prediction, and frame interpolation, and introduces Autoregressive Unrolling and Adaptive Context Switching to mitigate temporal drift in extremely long sequences.