论文

提出 SkillHone，一个基于持久决策历史实现 Agent Skill 持续进化的 harness；在内部工具辅助的分析场景中提升了准确率，并在未预先集成搜索栈的情况下优于商业支持的深度研究 Agent。SkillHone is introduced, a harness for continual agent skill evolution grounded in persistent decision history that improves accuracy on internal tool-mediated analysis scenarios and outperforms commercially backed deep-research agents without a pre-integrated search stack.

提出 conversational infill：让一个小型 talker 模型在外部 reasoner 模型产生结果前即时生成上下文相关的回复以掩盖延迟，并在推理过程中将 reasoner 流式输出的知识流畅地融合到回复中。Conversational infill is introduced, where a small talker model both immediately generates contextually grounded responses to hide the latency of an external reasoner model and fluently integrates streamed reasoner knowledge into its responses during inference.

AOHP 的核心设计原则是将 Agent 视为 OS 中的一等公民，从而支持自适应用户界面以及对 Agent 友好的运行时环境；在任务完成度、执行成本和安全策略合规性方面均展现出明显优势。The core design principle of AOHP is to treat agents as first-class OS actors, enabling adaptive user interfaces and agent-friendly runtime environments, and shows clear advantages in task completion, execution cost, and security-policy compliance.

本研究收集了使用主流 AI Agent 开发的大量真实应用语料，并设计了结合 Agent 辅助代码审计与人工验证的漏洞分析框架，揭示出与传统软件开发流程中常见的漏洞模式不同的反复出现的漏洞规律。This study collects a large corpus of real-world applications developed using popular AI agents and designs a vulnerability analysis framework that combines agent-assisted code auditing with human validation, and reveals recurring vulnerability patterns that differ from those commonly observed in conventional software development workflows.

概述了通过 LLM 进行空间推理所面临的挑战，并展望了搜索引擎与 LLM 集成、通过图增强推理来回答复杂空间问题的未来。The challenges associated with spatial reasoning through LLMs are outlined and a future in which search engines integrate with LLMs to answer complex spatial questions through graph-enhanced reasoning is envisioned.

HAKARI-Bench 是一个轻量级基准，将现有检索套件重建为统一格式的小型数据集（Nano-sets），支持在同一条件下对五类检索方法及其效率变体进行与模型无关的对比。HAKARI-Bench is a lightweight benchmark that reconstructs existing retrieval suites into small datasets (Nano-sets) in a unified format, enabling same-condition, model-agnostic comparison of five retrieval families and their efficiency variants.

本文命名了"检索状态锁定"这一失败模式，通过分离单一置信度分数所混淆的三个对象——答案表面、检索到的证据以及检索状态本身——来诊断该问题，并直接衡量"一致性盲区"。This work names the failure retrieval-state lock-in and diagnose it by separating the three objects a single confidence score conflates: the answer surface, the retrieved evidence, and the retrieval state itself, and measures the agreement blind spot directly.

本文提出 Randomized YaRN，一种通过将基于 YaRN 的位置外推与随机位置编码和长度课程相结合来提升长度泛化能力的训练方法，表明渐进式地将模型暴露于分布外位置分布是实现可泛化长上下文推理的有效方案。Randomized YaRN is proposed, a training method that improves length generalization by combining YaRN-based positional extrapolation with randomized positional encoding and a length curriculum, and suggests that progressively exposing models to OOD positional distributions provides an effective recipe for generalizable long-context reasoning.

本文提出新框架 Diffusion-LLM，将条件扩散模型集成到基于 LLM 的预测流水线中，展示了分布感知正则化在提升时间序列 LLM 的鲁棒性与泛化能力方面的价值。This work proposes a new framework Diffusion-LLM that integrates a conditional diffusion model into an LLM-based forecasting pipeline, and demonstrates the value of distribution-aware regularization for enhancing robustness and generalization in time series LLMs.

一个包含 382 个真实企业任务、覆盖 6 种专业角色和 22 项程序性技能的基准，用于评估技能在任务、角色和模型骨干间的迁移能力，发现部分技能可在任务和模型间广泛泛化，而另一些则专化为角色特定工作流，在迁移时失去效力。A benchmark of 382 realistic enterprise tasks spanning six professional roles and 22 procedural skills, designed to evaluate how skills transfer across tasks, roles, and model backbones finds that some skills generalize broadly across tasks and models, whereas others become specialized to role-specific workflows and lose effectiveness under transfer.

本文提出 RaysUp，一个超轻量级、任务无关且与 VFM 无关的特征上采样框架，可在任意分辨率下重建高分辨率特征图，仅使用 AnyUp 16% 的参数即达到 SOTA 性能，推理速度提升约 7 倍。RaysUp is proposed, an ultra-lightweight, task-agnostic, and VFM-agnostic feature upsampling framework that reconstructs high-resolution feature maps at arbitrary resolutions that achieves state-of-the-art performance while using only 16% of the parameters of AnyUp and delivering approximately 7x faster inference.

ABACUS 是一个统一视觉语言模型，可在无需任何基准特定训练的情况下处理物体计数、人群计数、指代表达计数以及忠实计数的图像生成，性能超越任务特定的专家模型和更大的通用模型。ABACUS is a unified vision-language model that handles object counting, crowd counting, referring-expression counting, and count-faithful image generation without any benchmark-specific training required, outperforming both task-specific specialists and larger generalist models.

本文提出一种将 LLM 与检索增强生成（RAG）相结合的混合方法，用于自动化跨版本 Qiskit 代码迁移，验证了这种以数据为中心的方法在促进技术独立性和提供缓解 API 过时问题的鲁棒智能助手方面的潜力。A hybrid approach integrating LLMs with Retrieval-Augmented Generation (RAG) to automate the migration of Qiskit code across versions and confirms the potential of this data-centric methodology to foster technological independence and provide robust, intelligent assistants that mitigate API obsolescence.

MedRLM 旨在将医疗 AI 从静态问答转向可审计、多模态且工作流感知的临床决策支持，并引入临床证据图记忆，将患者特定观察与检索到的证据相连接。MedRLM aims to move medical AI from static question answering toward auditable, multimodal, and workflow-aware clinical decision support, and introduces a Clinical Evidence Graph Memory to connect patient-specific observations with retrieved evidence.

本文命名并测量了推测性查询的检索收敛到包含答案结果时的输入流位置——"工具意图稳定化"：即推测性查询的检索收敛到包含答案结果的输入流位置。This work names and measures the point in the input stream at which a speculative query's retrieval converges on the answer-bearing result, tool-intent stabilization: the point in the input stream at which a speculative query's retrieval converges on the answer-bearing result.

对话式与工具使用的 LLM Agent 在上下文窗口中同时从多个方向被填充，而必须在多轮之间回忆信息的 Agent（即 memory 的典型场景）恰恰是 recency 截断失效的地方。Conversational and tool-using LLM agents operate over a context window that fills from several directions simultaneously, and agents that must recall information across many turns, the defining case for memory, are precisely where recency truncation fails.

提出 SAC，这是首个针对稀疏注意力模型优化的高效解耦 KV Cache 系统，利用 CXL（Compute Express Link）的低延迟、cache-line 粒度 load/store 语义，将基于 CXL 的解耦确立为新兴稀疏注意力模型的优越基础设施。SAC is proposed, the first efficient disaggregated KV cache system optimized for sparse attention models, which leverages the low-latency, cache-line granularity load/store semantics of Compute Express Link (CXL), establishing CXL-based disaggregation as the superior infrastructure for emerging sparse attention models.

提出 S-Agent，一种面向连续多视图图像与视频理解与推理的空间工具使用 Agent 范式，以无需训练的方式持续提升开源与闭源 VLM。This work introduces S-Agent, a spatial tool-use agentic paradigm for understanding and reasoning over continuous multi-view images and videos, and consistently improves both open-source and closed-source VLMs in a training-free manner.

揭示了指导的生成方式才是决定性变量，并提出 probe-and-refine tuning（探测-微调）：通过合成 bug 修复探测任务，利用单次 LLM 调用迭代诊断并修补仓库的指导文件，调优过程中不涉及 Agent 循环或工具调用。It is shown that how the guidance is produced is the decisive variable, and probe-and-refine tuning is introduced, a procedure that uses synthetic bug-fix probes to iteratively diagnose and patch a repository's guidance file through single-shot LLM calls, with no agent loop or tool use during tuning.

提出一种特权感知的训练后防御方法，教导 Agent 优先选用足够的低权限工具，仅在必要时升级；该方法在保留通用能力的同时大幅减少了不必要的高权限工具使用。A privilege-aware post-training defense that teaches agents to prefer sufficient lower-privilege tools and escalate only when necessary is introduced, showing that this defense substantially reduces unnecessary high-privilege tool use while preserving general capabilities.

提出 Vesta，一个统一的具身通用模型，将定位、空间推理、导航和长程规划能力整合到单个基础模型中，并证明通用模型能够达到或超越专家模型。Vesta is presented, a unified embodied generalist that consolidates localization, spatial reasoning, navigation, navigation, and long-horizon planning capabilities into a single foundation model and demonstrates that a generalist model can match or exceed specialists.

提出 Qwen-RobotNav，一种基于 Qwen-RobotNav 构建的可扩展导航模型，通过参数化接口沿两个互补维度应对该问题：用于选择导航行为的多种任务模式，以及用于控制视觉历史编码方式的可控观测参数。Qwen-RobotNav is presented, a scalable navigation model built on Qwen-RobotNav that addresses it through a parameterised interface with two complementary dimensions: multiple task modes that select the navigation behaviour, and controllable observation parameters that govern how visual history is encoded.

提出 CogniRoute，一种面向社交全模态推理的 schema 引导 Mixture-of-Experts（混合专家）框架，并引入路由感知强化学习，通过答案正确性、模态一致性推理与认知时序锚定等奖励联合优化 token 生成与专家分配。CogniRoute, a schema-guided Mixture-of-Experts framework for social omni reasoning, is introduced and route-aware reinforcement learning is introduced, which jointly optimizes token generation and expert allocation using rewards for answer correctness, modality-consistent reasoning, and cognitive temporal grounding.

HACD-H 为建模自适应人-AI 社交交互与开发社交智能 AI 系统提供了统一的理论基础，并表明社交智能源自长期社交认知的协同进化，而非孤立的对话能力。The HACD-H provides a unified theoretical foundation for modeling adaptive human-AI social interaction and developing socially intelligent AI systems and suggests that social intelligence emerges from long-term social cognitive coevolution rather than isolated conversational capabilities.

提出 GateMem，一个面向多主体共享内存 Agent 的基准，联合评估合法长程请求及其状态更新的效用、跨上下文授权边界的访问控制，以及 Agent 在收到显式删除请求后的主动遗忘能力。GateMem is introduced, a benchmark for multi-principal shared-memory agents that jointly evaluates utility for legitimate long-horizon requests with state updates, access control across contextual authorization boundaries, and agent-facing active forgetting after explicit deletion requests.

C-Trace 提出一个验证框架，将 GDPR 的部分要求（包括同意、目的限制、数据最小化与被遗忘权）表达为针对 Agent 执行轨迹的正式策略谓词。C-Trace (Compliance Trace based Runtime Agent Conformance Enforcement), a verification framework that expresses a subset of GDPR requirements, including consent, purpose limitation, data minimization, and the right to erasure, as formal policy predicates over agent execution traces is proposed.

RODS（Reward-driven Online Data Synthesis）通过将进度奖励方差重新用作零成本边界检测器，在 RL 训练与数据生成之间形成闭环，无需在训练已有的 rollout 之外增加额外推理。RODS (Reward-driven Online Data Synthesis) closes the loop between RL training and data generation by repurposing the progress reward variance as a practical, zero-cost boundary detector that requires no extra inference beyond the rollouts already computed for training.

对涵盖前沿闭源与开源模型、共 22 个模型在多个规模上的评估发现，所有模型家族均存在不可忽略的隐私泄露，且指令遵循能力与泄露率呈正相关。Evaluating 22 models spanning frontier proprietary and open-source models at multiple scales, it is found that all model families exhibit non-trivial leakage, and that instruction- following ability correlates with leakage rate.

结果表明仅评估求解器或仅评估答案是不足的：Agent 的差异不仅体现在发现关键 contingency 上，还体现在验证预算的使用、显式提交、类型强制、重复验证、基于证据的报告以及缓解行为等方面。The results show why solver-only or answer-only evaluation is insufficient: agents are distinguished not only by top-contingency discovery, but also by validation-budget use, explicit submission, type coercions, duplicate validations, evidence-backed reporting, and mitigation behavior.

Qwen-RobotManip 在所有 OOD 场景下大幅超越包括 π0.5 在内的已有 SOTA 模型，在 RoboChallenge 中排名第一，相对改进 20%，并在 AgileX ALOHA、Franka、UR、ARX 等真实机器人平台上完成验证。Qwen-RobotManip substantially outperforms prior state-of-the-art models, including $\pi$0.5, across all OOD settings, ranks 1st in RoboChallenge with a 20% relative improvement, and is validated on real-robot platforms including AgileX ALOHA, Franka, UR, and ARX.

提出一种基于物体中心的残差强化学习框架，利用物体位姿精化 VLA 动作，使观测空间紧凑，在仿真与现实之间能够一致迁移。An object-centric residual RL framework is proposed that refines VLA actions using object poses, enabling a compact observation space that transfers consistently between simulation and reality.

该工作以模拟的共享工作空间人机团队为受控实验环境，研究协作结构如何影响团队行为，并表明协调结构是决定可用能力能否提升团队结果的关键。This work uses simulated shared-workspace human-AI teams as a controlled testbed for studying how collaboration structure shapes team behavior, and suggests that coordination structure is central to whether available capability improves team outcomes.

提出 MCompassRAG，一种由元数据引导的检索框架，将主题级信号作为语义罗盘以选择相关证据，在同一 embedding 空间中以主题元数据丰富 chunk 表示，并通过 LLM 教师蒸馏训练轻量级检索器。MCompassRAG is introduced, a metadata-guided retrieval framework that uses topic-level signals as a semantic compass for selecting relevant evidence and enriches chunk representations with topic metadata in the same embedding space and trains a lightweight retriever through LLM-teacher distillation.

提出 HistoRAG，一个将史学原则转化为具体架构干预的框架，为特定领域认识论承诺如何转化为 RAG 设计决策提供模型，并可迁移至其他处理大规模语料的诠释性学科。HistoRAG is introduced, a framework that translates historiographical principles into concrete architectural interventions and offers a model for how domain-specific epistemological commitments can be translated into RAG design decisions, and may transfer to other interpretive disciplines working with large corpora.

实现了一个 RAG 框架，包含面向 LLM 的错误追踪与反例生成以提供更优上下文，并兼容最新版 Isabelle 与 Sledgehammer 以提升效率。A Retrieval-Augmented Generation framework, Error tracing and counterexample generation for improved context supplied to the Large Language Model, and Compatibility with the latest version of Isabelle and Sledgehammer is implemented for improved efficiency.

提出 HyGRAG，一种分层图 RAG 框架，通过解决三个核心挑战超越源文档限制：构建真正融合上下文与关系信息的摘要、利用这些综合表示在检索阶段访问涌现知识、以及为动态语料高效更新分层结构。HyGRAG is proposed, a hierarchical graph RAG framework that transcends source documents by addressing three core challenges: constructing summaries that genuinely integrate contextual and relational information, leveraging these synthesized representations to access emergent knowledge during retrieval, and efficiently updating hierarchical structures for dynamic corpora.