Python · ★ 12,509 · 🍴 811 · 📈 3,142 stars today
Compress tool outputs, logs, files, and RAG chunks before they reach the LLM. 60-95% fewer tokens, same answers. Library, proxy, MCP server.
中文介绍 这是一款工具库、代理和 MCP 服务器，专注于在数据进入大语言模型前对其进行压缩，涵盖工具输出、日志、文件及 RAG 块。通过预处理技术，在保持答案质量不变的前提下，可减少 60%-95% 的 token 消耗，适用于需要优化成本与效率的 LLM 应用开发者。
Python · ★ 180,998 · 🍴 31,053 · 📈 1,913 stars today
The agent that grows with you
中文介绍 该项目旨在构建一个能够伴随用户共同成长、自适应进化的 AI 代理。其核心概念聚焦于代理的持续学习与个性化发展，可能采用强化学习或反馈机制，适用于追求长期人机协作与个性化助手服务的开发者与研究者。
JavaScript · ★ 207,238 · 🍴 31,818 · 📈 1,750 stars today
The agent harness performance optimization system. Skills, instincts, memory, security, and research-first development for Claude Code, Codex, Opencode, Cursor and beyond.
中文介绍 该系统是面向 AI 代码代理（如 Claude Code、Codex 等）的性能优化框架，整合了技能、本能、记忆、安全等模块，并以研究为导向进行开发。它旨在提升代理的编码与任务执行能力，主要服务于使用这些 AI 编码工具以提高开发效率的软件工程师。
Python · ★ 79,871 · 🍴 10,601 · 📈 141 stars today
Turn any PDF or image document into structured data for your AI. A powerful, lightweight OCR toolkit that bridges the gap between images/PDFs and LLMs. Supports 100+ languages.
中文介绍 一套强大且轻量的 OCR 工具包，能够将任意 PDF 或图像文档转换为结构化数据，以供 AI 模型处理。它支持超过 100 种语言，有效弥合了非结构化图像/PDF 与 LLM 之间的鸿沟，适用于需要文档数字化与智能解析的数据科学家和开发者。
Python · ★ 108,573 · 🍴 9,598 · 📈 321 stars today
💫 Toolkit to help you get started with Spec-Driven Development
中文介绍 这是 GitHub 推出的一个工具包，旨在帮助开发者快速入门规范驱动开发（Spec-Driven Development）。它可能提供模板、最佳实践或自动化工具，以提升代码质量与开发效率，面向希望采用更严格、更清晰开发流程的工程团队。
Jupyter Notebook · ★ 8,998 · 🍴 580 · 📈 133 stars today
NVIDIA Cosmos is an open platform of world models, datasets, and tools that enables developers to build Physical AI for robots, autonomous vehicles, smart infrastructure, and more.
中文介绍 NVIDIA Cosmos 是一个开放的物理 AI 开发平台，提供世界模型、数据集及相关工具。它旨在帮助开发者为机器人、自动驾驶车辆、智能基础设施等应用构建物理世界仿真与决策能力，主要面向从事具身智能与自动化系统研发的工程师和研究员。
TypeScript · ★ 25,030 · 🍴 2,917 · 📈 212 stars today
An Open Source implementation of Notebook LM with more flexibility and features
中文介绍 一个开源版本的 Notebook LM 实现，在原有概念基础上提供了更高的灵活性和更丰富的功能。它旨在帮助用户更便捷地管理、整合与查询知识，适用于研究人员、学生和知识工作者，用于构建个人或团队的第二大脑系统。
Python · ★ 9,590 · 🍴 1,153 · 📈 581 stars today
Talk to any LLM with hands-free voice interaction, voice interruption, and Live2D taking face running locally across platforms
中文介绍 该项目实现了一套完全本地运行的免提交互系统，允许用户通过语音与任何大语言模型对话，支持语音打断和 Live2D 虚拟形象驱动。它整合了语音识别、合成与模型推理技术，适用于内容创作者、VTuber 或希望以更自然方式与 AI 交互的开发者。
★ 349,729 · 🍴 83,226 · 📈 632 stars today
A complete computer science study plan to become a software engineer.
中文介绍 一个完整的计算机科学学习计划，旨在系统性地帮助学习者掌握必要知识与技能，以成功通过软件工程师的面试。它涵盖了从数据结构、算法到系统设计等多个核心领域，主要面向正在准备技术面试的求职者和转行者。
Java · ★ 8,976 · 🍴 1,211 · 📈 38 stars today
Multi-platform SDK for integrating GitHub Copilot Agent into apps and services
中文介绍 这是一个多平台软件开发工具包（SDK），用于将 GitHub Copilot Agent 的能力集成到各类应用程序和服务中。它为开发者提供了一套标准化的接口和工具，便于在第三方产品中嵌入 AI 辅助编程或智能体功能。
Go · ★ 35,668 · 🍴 430 · 📈 255 stars today
Find vulnerabilities, misconfigurations, secrets, SBOM in containers, Kubernetes, code repositories, clouds and more
中文介绍 一款全面的安全扫描工具，能够检测容器、Kubernetes、代码仓库、云环境等多种目标中的漏洞、配置错误、敏感信息泄露及软件物料清单（SBOM）。它广泛用于 DevSecOps 流程，帮助安全工程师和开发者在早期发现并修复风险。
C# · ★ 1,326 · 🍴 169 · 📈 411 stars today
Windows companion suite for OpenClaw - System Tray app, Shared library, Node, and PowerToys Command Palette extension
中文介绍 这是 OpenClaw 项目的 Windows 伴侣套件，包含系统托盘应用、共享库、节点以及 PowerToys 命令面板扩展。它旨在为 Windows 用户提供与 OpenClaw 主服务无缝集成的本地化体验，增强系统的可访问性与易用性。
TypeScript · ★ 5,309 · 🍴 640 · 📈 308 stars today
A modern platform for visual, flexible, and extensible graph-based investigations. For cybersecurity analysts and investigators.
中文介绍 一个现代化的可视化调查平台，基于图模型构建，支持灵活、可扩展的关联分析。它专为网络安全分析师和调查人员设计，用于对事件、威胁指标等进行可视化追踪与关系梳理，以支持复杂的威胁狩猎与事件响应工作。
Python · ★ 27,579 · 🍴 2,345 · 📈 199 stars today
AI agent skill that researches any topic across Reddit, X, YouTube, HN, Polymarket, and the web - then synthesizes a grounded summary
中文介绍 这是一个 AI 代理技能，能够跨越 Reddit、X、YouTube、Hacker News、Polymarket 及网络等多源平台，对指定话题进行研究，并最终合成一个有据可依的摘要报告。它适用于研究人员、投资者或任何需要快速概览特定领域近期动态的用户。
@elpresidank · 116 粉丝 · 2.9M 阅 · 543 赞 · 35 转
Most AI agent memory is built on embeddings. And there's now a proof that this entire class of system is going to forget what you stored in it — and confidently make up things you never stored at all.
中文介绍 AI代理的向量记忆系统存在根本缺陷，有理论证明基于嵌入的存储会遗忘信息并产生幻觉。博主主张用拓扑结构替代纯嵌入来构建代理的长期记忆，以解决遗忘与混淆问题。
@DamiDefi · 96.5K 粉丝 · 2.3M 阅 · 584 赞 · 80 转
The number that stopped me was not the $2 trillion valuation. It was $791 million. That is what SpaceX made in net income in 2024. A profitable, growing aerospace company with a genuine moat in launch
中文介绍 博主将SpaceX长达300页的S1招股文件输入Claude进行分析，发现了隐藏的关键财务数据：2024年净收入达7.91亿美元。这展示了AI在快速解析复杂商业文档、提取核心信息方面的强大能力。
@0xCodez · 3.3K 粉丝 · 637.2K 阅 · 510 赞 · 59 转
Most Claude Code users still write their workflows by hand. They chain prompts, copy outputs, paste them into the next prompt, fix what went wrong, repeat. 9 out of 10 builders haven’t tried Dynamic
中文介绍 一份针对Claude Code用户的进阶教程，介绍了Anthropic工程师实际使用的6种动态工作流模式和14个具体步骤，旨在让工作流由AI动态生成，而非手动编写和拼接。
@Saboo_Shubham_ · 116.2K 粉丝 · 263.3K 阅 · 517 赞 · 74 转
The frontend used to be a fixed thing. Designers drew it. Engineers built it. Users got what shipped. That's over. The interfaces shipping in 2026 are drawn partly by the agent itself, in real time,
中文介绍 预测前端开发将进入“生成式UI”时代，界面不再完全由设计师预先固定，而是由AI代理在运行时根据上下文实时生成和绘制，这将彻底改变应用的交互方式。
@mem0ai · 17.6K 粉丝 · 82.8K 阅 · 520 赞 · 60 转
Agent harnesses are where AI software actually runs. Cursor, Devin, Claude Code, Codex: these environments handle context, orchestrate tools, coordinate agents, and increasingly, manage memory. The
中文介绍 分析了主流AI代理运行环境（如Cursor、Devin、Claude Code）的现状，指出这些“代理缰绳”核心功能是管理上下文、编排工具与协调代理，而记忆管理正成为其演进的关键方向。
@trq212 · 263.1K 粉丝 · 75.7K 阅 · 542 赞 · 36 转
Last week, we released dynamic workflows in Claude Code. Claude can now write its own harness on the fly, custom-built for the task at hand. While the default Claude Code harness is built for coding,
中文介绍 介绍了Claude Code新推出的“动态工作流”功能：Claude能够根据当前任务，实时编写和生成一个完全定制化的运行环境，超越了预设的静态工作流，提升了处理复杂任务的灵活性。
@drfeifei · 738.0K 粉丝 · 72.2K 阅 · 699 赞 · 144 转
“The world is everything that is the case.” — Ludwig Wittgenstein, Tractatus Logico-Philosophicus, 1921 The world is not made of words. In an earlier essay, we argued that spatial intelligence is AI’s
中文介绍 李飞飞团队发布论文，对“世界模型”提出了一个功能性分类体系。文章回顾了早期的空间智能观点，旨在超越语言模型，系统性理解AI如何构建对物理世界的认知与建模。
@sydneyrunkle · 7.5K 粉丝 · 69.5K 阅 · 511 赞 · 74 转
Building useful agents is largely about customization: connecting your agent to the right context, data, and environment(s) for the task at hand. At its core, an agent is a model calling tools in a
中文介绍 一篇关于构建自定义AI代理运行环境（Harness）的指南，强调代理的核心是模型在特定上下文和数据中调用工具，而定制化是构建实用代理的关键，需要连接正确的环境。
@mvanhorn · 27.6K 粉丝 · 54.5K 阅 · 545 赞 · 44 转
Three months ago I posted "Every Claude Code Hack I Know." It hit 913K views. @kevinrose had asked what IDE to use, and my answer was: "No IDE. Just plan.md files and voice." This used to be called
中文介绍 博主汇总了其已知的所有代理工程技巧，作为早先“Claude Code技巧”帖子的更新版。其核心方法是抛弃IDE，改用计划文件（plan.md）和语音来驱动Claude完成复杂任务。
@jainarvind · 9.3K 粉丝 · 53.7K 阅 · 505 赞 · 68 转
Enterprise AI token spend is scaling quickly, especially as the technology shifts from simple chat assistants into coding agents, AI coworkers, and long-running workflows. These systems do far more
中文介绍 指出企业AI的代币消耗激增，不仅是模型本身的问题，更是AI架构问题。随着代理从聊天助手转向长时运行的工作流，需要从系统层面优化架构来控制成本。
@IBuzovskyi · 1.2K 粉丝 · 50.1K 阅 · 500 赞 · 51 转
These 10 Hermes Agent hacks saved me 15+ hours every week - and they work for any workflow you run repeatedly. Content, software development, business operations, client management, research, sales.
中文介绍 分享了10个Hermes Agent的实用技巧，据称每周能节省15小时以上。这些技巧适用于内容创作、软件开发、业务运营等多种重复性工作流程，旨在将聊天代理转变为自动化系统。
@delba_oliveira · 74.0K 粉丝 · 37.4K 阅 · 533 赞 · 38 转
As we delegate more ambitious tasks to Claude, it becomes increasingly important that it can verify its own work. The more Claude can self-verify: the more independently it can work on long-running
中文介绍 提出通过建立“反馈循环”来增强Claude Code的自主性。让AI能够自我验证工作成果，从而减少人工监督，使其能更独立地处理复杂的长时任务。
@ericzakariasson · 67.9K 粉丝 · 37.3K 阅 · 507 赞 · 27 转
If you've ever watched an agent try to fix a bug, you've watched it guess. It reads the code, comes up with a theory, makes an edit, and hopes. Sometimes it's right. A lot of the time you get a fix
中文介绍 建议为AI代理提供实时的运行时上下文（如状态、日志、环境变量），而不是仅让其基于静态代码猜测。这能显著提升代理调试和修复Bug的准确性与效率。
@OptimaiNetwork · 98.8K 粉丝 · 5.7K 阅 · 542 赞 · 110 转
May was an important month for OptimAI because it marked a shift from proving the product to expanding the platform. The broader market is moving in the same direction: AI is no longer just about
中文介绍 OptimAI项目月度回顾，宣布重点从产品验证转向平台扩展。认为行业趋势正从单一的AI应用转向构建“代理原生”的互联网基础设施，AI将深度融入软件底层。
We talk with the VendingBench authors on evaling Claudes from Haiku to Mythos, and how they build leading, and lasting, frontier evals from scratch.
中文介绍 Andon Labs的Lukas Petersson和Axel Backlund讨论了如何评估Claude系列模型（从Haiku到Mythos版本），以及他们如何从零开始构建领先且持久的前沿评估基准。
中文介绍 NVIDIA发布了Nemotron 3.5内容安全模型，旨在为全球企业AI提供可定制的多模态安全解决方案。
中文介绍 EVA-Bench Data 2.0发布，涵盖3个领域、121种工具和213个场景，旨在为AI代理提供更全面的评估数据集。
Learn how Endava is using AI agents, ChatGPT Enterprise, and Codex to accelerate software delivery, automate workflows, and build an AI-native culture across the enterprise.
中文介绍 Endava正在利用AI代理、ChatGPT企业版和Codex来重新设计软件交付流程，以加速开发、自动化工作流程并在企业内部构建AI原生文化。
Most days in her chambers, Judge Maritza Braswell, a federal magistrate judge in Colorado, sifts through stacks of documents written by people without a lawyer. Many of them can’t afford to hire a lawyer, and others have cases too weak or too small to interest one. She reads each one carefully, mind
中文介绍 法院正面临AI生成诉讼的激增。例如，科罗拉多州联邦治安法官Maritza Braswell需处理大量由无力聘请律师的当事人提交的文件。
ChatGPT introduces a new memory system to better remember preferences, keeping context fresh and relevant across conversations.
中文介绍 ChatGPT推出名为“Dreaming”的新记忆系统，旨在更好地记住用户偏好，使对话间的上下文保持新鲜和相关。
a quiet day.
中文介绍 新闻提及图像生成领域的新进展，包括Reve 2和Ideogram 4在布局方面的能力，今日该领域相对平静。
An action plan for AI-powered biological resilience
中文介绍 OpenAI发布了一份关于情报时代生物防御的行动计划，旨在利用人工智能增强生物韧性。
中文介绍 Hugging Face博客讨论了将hf CLI设计为一种优化的代理方式，以便与Hub进行交互和工作。
中文介绍 AI日报提及：DeepSeek正在进行融资，Meta的模型发布延迟，以及Gemma 4 12B模型的相关动态。
Verified Generation and Compounding Intelligence
中文介绍 Axiom Math的Carina Hong探讨了如何超越非正式人工智能，重点关注验证生成和复合智能的规模化发展。
The legendary Microsoft CEO makes his first Latent Space appearance!
中文介绍 微软首席执行官Satya Nadella首次做客Latent Space，在Microsoft Build大会上与No Priors进行了一场特别访谈。
GPT-Rosalind advances life sciences research with enhanced biological reasoning, medicinal chemistry expertise, genomics analysis, and experimental workflow capabilities.
中文介绍 OpenAI为GPT-Rosalind引入新功能，增强了其生物推理、药物化学、基因组学分析和实验工作流程等方面的能力，以推进生命科学研究。
中文介绍 Dharma AI的博客探讨了直接偏好优化（DPO）技术在聊天机器人之外的更广泛应用前景。
See how Wasmer used Codex with GPT-5.5 to build a Node.js runtime for the edge, accelerating development 10x to 20x and shipping in weeks instead of months.
中文介绍 Wasmer利用Codex与GPT-5.5构建了面向边缘计算的Node.js运行时，将开发速度提升了10到20倍，并将交付时间从数月缩短至数周。
第一作者: Jian Yang · 方向: 密码学协议
数据隐私因果结构学习全同态加密计算优化
Abstract:Preserving data privacy is an important topic in structural data management and data mining. However, the issue of privacy leakage in distributed causal structure learning is a persistent challenge, especially in cases where data transmission and computation are required. In this paper, we propose a method based on fully homomorphic encryption (FHE) that performs calculations on ciphertexts, keeping data encrypted in transition and computation. Nevertheless, adopting FHE to causal structure learning is challenging due to the high computation cost and limited support on division as well as logarithm operations in FHE. To tackle this challenge, we propose a series of novel techniques including (i) circuit simplification for better efficiency, (ii) approximation of division and logarithm through Newton-Raphson Reciprocal and Taylor expansion, and (iii) a batching technique with...
论文介绍 本文研究分布式因果结构学习中的数据隐私泄露问题。提出基于全同态加密（FHE）的方法，在密文上执行计算，保持数据在传输和计算中加密。针对FHE计算成本高和对除法、对数支持有限的挑战，创新性地引入电路简化、除法和对数近似（如牛顿-拉夫逊倒数和泰勒展开）等技术，以提高效率。该方法为隐私保护下的因果结构学习提供了可行方案。
第一作者: Mohammed Gharib · 方向: 系统安全
活体检测被动检测神经肌肉微运动IMU传感器系统安全
Abstract:Liveness detection has evolved from a safeguard against presentation and replay attacks in biometric authentication to a broader requirement for distinguishing human users from non-human agents in modern digital systems. The emergence of generative and agentic AI further amplifies this need, positioning liveness as a fundamental security primitive. Existing approaches face key limitations, including reliance on explicit user interaction, specialized hardware, vulnerability to increasingly realistic spoofing, and limited scalability in real-world deployments. We present A-Live, a passive liveness detection framework that operates solely on inertial measurement unit (IMU) signals available in commodity devices. A-Live is based on the observation that neuromuscular micro-motions inherent to human motor control produce subtle but measurable signatures in inertial data, which are...
论文介绍 本文提出A-Live，一个被动活体检测框架，仅利用商品设备中的惯性测量单元（IMU）信号。基于人类运动控制中固有的神经肌肉微运动在惯性数据中产生可测签名的观察，A-Live无需用户交互或专用硬件即可检测活体。该方法解决了现有方法的限制，如依赖交互、硬件需求和易受欺骗，为可扩展的活体检测提供新途径。
第一作者: Talal Ashraf Butt · 方向: 软件安全
信任欺诈稀疏评级网络贝叶斯最优伯努利CUSUM序列检测
Abstract:Sequential trust detection in rating networks relies on continuous observation models that fail on real data. On Bitcoin-OTC, 56\% of ratings take a single value under standard mapping, breaking the distributional assumptions that parametric detectors require. This paper makes three contributions. It derives a Bayes-optimal F1 detection ceiling for per-node sequential detectors using empirically measured observation parameters. At Bitcoin-OTC's median in-degree of 2, this ceiling falls to 0.451 for strategic attacks, explaining why unsupervised methods cluster near $F1 \approx 0.4$. The analysis shows that detector-model matching, not information content, determines performance: binary models retain 86\% of mutual information while enabling exact parametric fit. A dual-regime architecture is presented where Bernoulli CUSUM detects behavioral shifts and triggers asymmetric...
论文介绍 本文研究稀疏评级网络中的信任欺诈检测。现有连续观察模型在真实数据上失效，如Bitcoin-OTC数据中56%的评级取单值。提出贝叶斯最优F1检测上限，在中位入度为2时上限为0.451。基于此，设计双regime架构，使用伯努利CUSUM检测行为变化并触发不对称机制，以提升检测性能。
第一作者: Micaella Bruton · 方向: 软件安全
同音替换密码自动破译注意力LSTM密码分析历史文本
Abstract:Homophonic substitution ciphers replace each plaintext letter with one of several possible ciphertext codes, deliberately weakening letter-frequency patterns and making automated decipherment difficult. This paper evaluates whether an attention-augmented Long Short-Term Memory (LSTM) model can learn such mappings in a historically motivated shared-key setting: all ciphertexts draw from the same known homophonic code pool, while individual keys use different consistent subsets of that pool. Using synthetic ciphertexts generated with ChronoFidelius from historical English and Swedish texts dated 1500--1899, we test performance across ciphertext lengths, centuries, variable-length codes, and simulated transcription errors. Models are trained only on aligned ciphertext--plaintext pairs, without external language models, frequency statistics, or key-search heuristics. Results show...
论文介绍 本文评估注意力增强LSTM模型在自动破译同音替换密码中的性能。在历史英语和瑞典语文本（1500-1899年）生成的合成密码上测试，模型仅使用对齐的密文-明文对训练，无外部语言模型或频率统计。实验考察了密码长度、世纪、变长代码和转录错误的影响，展示了模型在历史密码破译中的能力。
第一作者: Peihua Mai · 方向: AI 安全
隐私保护LLM推理模型无关批次优化提示安全
Abstract:With the widespread deployment of public large language models (LLMs) such as ChatGPT, protecting user prompt privacy has become an increasingly critical issue. Existing privacy-preserving inference methods sacrifice either utility or efficiency, and often require model-specific modifications that limit their compatibility. In this paper, we propose SharedRequest, a model-agnostic framework for privacy-preserving LLM inference that reformulates privacy protection at the batch level rather than the individual-prompt level. The key idea is to obscure sensitive information by mixing original prompts with noisy variants, while grouping semantically equivalent instructions to amortize the inference cost over a large batch of queries with minimal impact on LLM response quality. This design is independent of the LLM architecture, requiring no access to model parameters or...
论文介绍 本文提出SharedRequest，一个模型无关的隐私保护LLM推理框架。核心思想是在批次级别保护隐私，通过混合原始提示与噪声变体，并分组语义等效指令，以摊销推理成本。该设计无需访问模型参数或修改架构，在最小化对响应质量影响的同时，有效模糊敏感信息，适用于公共LLM如ChatGPT。
第一作者: Yiqi Wang · 方向: AI 安全
LLM代理证据追踪执行溯源可信AI审计
Abstract:Large language model (LLM)-based agents increasingly solve complex tasks by interacting with external tools, retrieval systems, memory modules, environments, and other agents. These capabilities expand agent autonomy, but also make agent behavior harder to verify, debug, and audit. Final-answer accuracy alone cannot explain how an output was produced, which evidence supported each claim, whether tool calls were justified, how memory influenced later decisions, or where execution failures originated. Evidence tracing and execution provenance address this gap by modeling how retrieved evidence, tool outputs, memory items, environment observations, intermediate claims, actions, and final answers are connected throughout agent execution. This survey provides a systematic review and conceptual framework for evidence tracing and execution provenance in LLM agents. We organize...
论文介绍 本文综述LLM代理中的证据追踪和执行溯源研究。随着LLM代理与外部工具、检索系统等交互，其行为验证变得困难。证据追踪建模证据、工具输出、内存等连接，执行溯源记录执行过程。文章提供系统性框架，帮助理解输出生成、证据支持和执行失败，增强LLM代理的可审计性和信任。
第一作者: Samuel Ndichu · 方向: 密码学协议
异常检测安全运营中心日志重写自然语言TreeSHAP
Abstract:System-generated logs underpin security monitoring, yet their rigid template-based format hinders both automated analysis and human comprehension. We present NLLog (Natural-Language Log), a lightweight pipeline that deterministically rewrites parsed templates into WHO-WHAT-SEVERITY sentences, pools them with term-frequency-inverse-document-frequency weighting, classifies sessions with tree ensembles, and back-projects evidence with TreeSHAP for analyst review. On Hadoop Distributed File System (HDFS) and Blue Gene/L (BGL) corpora, NLLog exceeds two reproduced matched-protocol baselines; across HDFS, BGL, and the AIT Alert Data Set, it sustains low false-positive rates with commodity-hardware latency suitable for security operations center triage. Coverage, sparse-versus-dense, faithfulness, and adversarial ablations show that fallback sufficiency is corpus-dependent, that an...
论文介绍 本文提出NLLog，一个轻量级、可解释的SOC异常检测管道。它将解析日志模板确定性重写为WHO-WHAT-SEVERITY自然语言句子，使用TF-IDF加权池化，并用树集成分类会话，最后用TreeSHAP背投证据。在HDFS、BGL和AIT数据集上，NLLog超越基线，保持低误报率，适合商品硬件上的SOC分诊。
第一作者: Pinshen Xu · 方向: 密码学协议
可信执行环境异构TEE分布式自治组织安全多方计算信任管理
Abstract:Trusted Execution Environments (TEEs) have emerged as a critical technology for safeguarding sensitive data and ensuring code integrity in modern computing systems. However, relying on a single TEE implementation makes systems vulnerable to a central point of attack. Building distributed-trust systems leveraging heterogeneous TEEs helps disperse trust but still faces threats from centralized management and adaptive mobile adversaries. To address these challenges, this paper introduces TeeDAO, a novel three-layer framework that automatically organizes multiple heterogeneous TEE instances and provides unified interfaces to support diverse applications, while ensuring long-term guarantees of availability, integrity, and confidentiality. TeeDAO couples BFT-ordered governance with heterogeneity-aware Distributed Proactive Secret Sharing (DPSS) and Secure Multi-Party Computation...
论文介绍 本文介绍TeeDAO，一个用于异构可信执行环境（TEEs）的分布式自治组织框架。针对单TEE的中心点攻击风险，TeeDAO提供三层架构，自动组织异构TEE实例，并耦合BFT治理与分布式主动秘密共享和安全多方计算。它确保可用性、完整性和机密性的长期保证，为分布式信任系统提供统一接口。
第一作者: Varun Kohli · 方向: 网络安全
LEO-ISL跨层指纹攻击检测卫星网络安全物理网络层融合
Abstract:Low-Earth Orbit (LEO) mega-constellations such as Starlink by SpaceX and Kuiper by Amazon rely on optical Inter-Satellite Links (ISLs) for autonomous mesh routing to provide low-latency telecommunication, Internet of Things (IoT), and security services globally. As commercial operators and governments deploy increasingly dense constellations and form multi-operator peering coalitions, ISL integrity becomes critical to both commercial availability and national security. However, there is a lack of real-world data for LEO constellations and existing real-time security approaches focus strictly on physical layer security, leaving blind spots in the coverage of network-layer and composite attacks. In this paper, we present a cross-layer, lightweight behavioral fingerprinting framework that fuses onboard physical-layer measurements with network-layer data to detect anomalies at low...
论文介绍 低地球轨道（LEO）星座依赖星间链路（ISL）实现全球通信，但现有安全方法仅关注物理层，导致网络层攻击检测存在盲点。本文提出 CLIF 框架，融合物理层测量和网络层数据，通过轻量级行为指纹技术检测物理和网络复合攻击。该研究有助于提升密集星座的安全性，适用于商业和国防通信领域。
第一作者: Guanlong Wu · 方向: 系统安全
联邦学习可信执行环境聚合安全分布式系统隐私保护
Abstract:Trusted Execution Environments (TEEs)-aided federated learning protocols emerge as promising solutions to counter server-side adversaries and ensure the trustworthiness of the server. In this paper, we dissect existing protocols and demonstrate that server-side adversaries can still manipulate client selection and replay aggregation to compromise system robustness and privacy, by exploiting TEE limitations, i.e., state rollback and I/O manipulation. To this end, we present DIST-FL, a distributed system of servers guarded by multiple TEEs forming an append-only ledger for privacy-preserved, robust FL aggregation. Specifically, DIST-FL ensures operation linearizability to thwart state rollback attacks and incorporates inputs from reliable servers to mitigate I/O manipulation threats. We implement DIST-FL and conduct evaluations in WAN settings. Experimental results demonstrate...
论文介绍 基于可信执行环境（TEE）的联邦学习协议存在服务器端攻击风险，如状态回滚和 I/O 操纵。本文提出 DIST-FL 系统，使用多个 TEE 构建仅追加账本，确保操作线性化以防御状态攻击，并引入可靠输入缓解 I/O 威胁。该系统在广域网环境中评估，增强了聚合过程的安全性和隐私保护。
第一作者: Ju Yang · 方向: 系统安全
机密区块链可信执行环境委托执行攻击防御保密性
Abstract:Confidential blockchains leveraging Trusted Execution Environments (TEEs) have garnered extensive attention for transaction confidentiality. In this paper, we first taxonomize two classes of attacks against confidential blockchains, i.e., execution-inference and execution-replay attacks, which exploit TEEs' long-lasting side-channel and state-continuity issues to compromise the confidentiality of existing consortium blockchains. Then, we present ODYSSEY, a confidential blockchain that efficiently mitigates these attacks. The core innovations of ODYSSEY are the following: (1) Its delegation model: clients delegate transaction execution to their designated trustees, while other participants synchronize only the execution results, which significantly reduces the attack surface while preserving confidentiality and system performance. (2) Two novel techniques to improve ODYSSEY's...
论文介绍 机密区块链面临执行推理和执行重放两类攻击，利用 TEE 的侧信道和状态连续性问题。本文提出 ODYSSEY 系统，通过委托执行模型，让客户端委托受托人执行交易，其他参与者仅同步结果，减少攻击面。采用新技术提升性能，重建区块链的保密性。
第一作者: Abhinaba Basu · 方向: 密码学协议
有用工作证明区块链AI推理实证研究GPU资源
Abstract:Pearl, a Layer-1 blockchain with high-profile AI industry endorsements, markets its Proof-of-Useful-Work (PoUW) protocol as simultaneously securing the network and performing AI inference. We present the first systematic empirical measurement of a deployed PoUW system, finding that Pearl's 24 EH/s network -- representing approximately 320,000 GPU-equivalents consuming an estimated 112 MW -- produces zero useful AI computation. Budget GPU rental prices rose 38% and utilization surged from 57% to 94% following the mining software's public release, displacing legitimate research workloads. Our measurements span five dimensions: (1) network composition analysis of 8,012 workers shows all have inference-capable hardware, yet the dominant mining software contains no inference code; (2) the verification protocol accepts random matrices by design, confirmed by 44 pool-accepted shares...
论文介绍 对 Pearl 的有用工作证明（PoUW）协议进行实证研究，发现其 24 EH/s 网络未产生任何有用的 AI 计算，尽管硬件支持推理。预算 GPU 租赁价格上升、利用率激增，影响合法研究工作。测量涵盖网络组成、验证协议等多维度，揭示系统存在有用性差距。
第一作者: Yutao Shi · 方向: 密码学协议
模型上下文协议大语言模型描述代码不一致安全分析分类学
Abstract:The Model Context Protocol (MCP) has emerged as a critical standard empowering Large Language Models (LLMs) to utilize external tools. In this ecosystem, LLMs rely on natural language descriptions provided by MCP servers to select and execute functions. This interaction implicitly assumes that tool descriptions faithfully reflect their underlying implementations, while this assumption is not mandatorily verified in practice. As a result, MCP deployments may suffer from a problem named Description-Code Inconsistency (DCI), where a tool's description of its capabilities and security boundaries is not consistent with what the code actually does. In this paper, we present a comprehensive study of DCI in real-world MCP servers. We formally define the problem and propose a comprehensive taxonomy spanning functionality inconsistencies and undeclared side effects. Guided by this...
论文介绍 模型上下文协议（MCP）使大语言模型（LLM）能调用外部工具，但工具描述与代码实现可能不一致。本文定义描述-代码不一致（DCI）问题，提出功能和副作用分类，并进行测量与检测。研究揭示了安全影响，有助于提升 MCP 部署的可靠性。
第一作者: Jianwei Tai · 方向: 安全研究
链式思维答案劫持安全诊断模型鲁棒性验证方法
Abstract:We study a controlled numeric proxy for chain-of-thought (CoT) answer hijacking, motivated by attacks in which benign-looking reasoning steers a harmful final answer. CoT wrappers on GSM8K and MATH-500 flip final answers away from gold labels. Rather than treating activation patching as clean-trace restoration, we ask where hijacked trajectories are fragile and whether recovery depends on a same-problem clean source. Across Qwen2.5-7B and Llama3-8B on GSM8K few-shot, puzzle, and sycophant hijacks, three few-shot/puzzle cells pass confirmatory $K{=}1$ localization after Bonferroni correction. A selection-aware 50/50 band validation preserves held-out in-band minus out-of-band gaps of +32.6, +45.1, and +17.7 points for Qwen-puzzle, Llama3-fewshot, and Llama3-puzzle, while exact $\Lstar$ agreement is much less stable. Qwen-fewshot remains exploratory, and sycophant cells are...
论文介绍 研究链式思维（CoT）答案劫持攻击，其中良性推理引导有害最终答案。本文提出选择感知诊断方法，分析劫持轨迹的脆弱性。通过在 GSM8K 和 MATH 数据集上的实验，评估恢复依赖于干净源，提升了模型鲁棒性的理解。
第一作者: R.D.N. Shakya · 方向: 密码学协议
后量子密码软件系统实施挑战人因分析系统知识
Abstract:The transition to Post-Quantum Cryptography (PQC) is essential to protect software systems from emerging quantum-enabled threats. Although standardised PQC algorithms are now available, developers and organisations continue to face significant challenges in integrating them into real-world software systems. While existing studies primarily focus on cryptographic performance and algorithmic security, it provides limited understanding of the broader socio-technological factors that influence successful PQC implementation. This SoK investigates PQC implementation approaches and challenges through the Human, Organisation, and Technology (HOT) dimensions. By systematically synthesising existing approaches across these dimensions, we reveal a notable imbalance in the current body of knowledge, where technological solutions dominate, while human and organisational considerations...
论文介绍 后量子密码（PQC）对软件系统安全至关重要，但实施面临挑战。本文通过人、组织和技术（HOT）维度进行系统知识研究，调查 PQC 实施方法。揭示当前知识不平衡，技术方案主导，而人因和组织因素被忽视，为实际部署提供指导。
第一作者: Roy Ricaldi · 方向: 网络安全
网络犯罪Telegram社区发现框架工具数据集
Abstract:This paper presents TeleHunt, a framework and tool for evaluating the effectiveness of different strategies to discover cybercriminal communities on Telegram. TeleHunt employs a set of reference-driven snowballing strategies, integrating message-level classification, contextual filtering, and market-segment labeling. Using open- and dark-web seeds, we systematically evaluate how seed source, pointer type, and exploration strategy influence discovery outcomes in three dimensions: efficiency, accessibility, and rediscovery. Our work provides (i) a modular cybercrime content discovery pipeline, (ii) the first systematic comparison of Telegram discovery strategies with an empirical characterization of market-segment accessibility, and (iii) a labeled dataset of over 172 million messages from 6,022 Telegram communities.
论文介绍 本文提出 TeleHunt 框架和工具，用于评估在 Telegram 上发现网络犯罪社区的策略。采用参考驱动的雪花球策略，整合消息分类、上下文过滤和市场标签。通过系统评估，提供模块化管道、策略比较和标记数据集，支持网络安全研究。
第一作者: Hiroki Nakano · 方向: AI 安全
威胁情报多智能体系统社交媒体网络安全IoC
Abstract:Cyber threat signals are fragmented across multiple social media platforms, yet no existing approach has fully automated their integration into actionable threat intelligence (TI) reports. We present TIBlender, a multi-agent system that monitors four platforms (X, Reddit, Telegram, and Discord) and produces structured TI reports via role-specialized LLM agents. These agents conduct multi-perspective investigations, tracing chains of evidence to uncover related Indicators of Compromise (IoCs) via collaborative, evidence-backed analysis. In a real-world deployment, TIBlender detected emerging threats across all four threat categories ahead of public feeds, including in-the-wild exploitation ahead of public vulnerability registries; the majority of its IoCs were absent from each evaluated feed. Quantitative evaluation confirms that each platform contributes unique threat...
论文介绍 该研究针对网络威胁信号跨平台分散、难以自动化整合为可操作情报的问题，提出了TIBlender多智能体系统。该系统监控X、Reddit等四个平台，通过角色分工的LLM智能体进行协作分析，追踪证据链并识别相关威胁指标。在实际部署中，它能够比公开情报更早地发现新兴威胁，其多数威胁指标未被现有情报源收录。该研究为自动化构建跨平台早期威胁情报提供了新方法。
第一作者: Jingkai Mao · 方向: 软件安全
机密虚拟机用户空间完整性特权分离完整性强制可执行对象
Abstract:Confidential Virtual Machines (CVMs), such as AMD SEV-SNP, enable cloud tenants to run security-sensitive workloads, but tenants can rely on the execution of these workloads only when they can trust the CVM. This trust requires continuous integrity assurance from CVM launch to the current runtime state, including initial trust establishment at launch and subsequent runtime integrity assurance. Existing works help establish launch-time trust and protect parts of runtime integrity, but they do not fully address the integrity of file-backed user-space executable objects, such as main executables, program interpreters, and dynamically loaded shared objects, that may be loaded or mapped dynamically during execution inside CVMs. In this paper, we propose Privilege-Separated User-space Integrity Enforcement (PS-UIE), an approach for enforcing the integrity of user-space executable...
论文介绍 机密虚拟机（如AMD SEV-SNP）需要持续的完整性保证来信任安全工作负载，但现有方法未完全覆盖文件支持的用户空间可执行对象的运行时完整性。本文提出特权分离用户空间完整性强制（PS-UIE）方法，通过特权分离机制强制执行这些对象的完整性。该方法可增强云环境中安全敏感工作负载的信任基础，确保从启动到运行时状态的连续完整性保护。
第一作者: Daniel Zhao · 方向: 安全研究
机密虚拟机完整性验证特权分离用户空间执行对象
Abstract:Watermarking methods for language models have been studied extensively in the autoregressive setting, where tokens are generated sequentially. These works largely focus on local-context schemes that perturb the next token's distribution as a function of its preceding tokens. In diffusion language models, distributions over many unresolved positions are jointly sampled, allowing additive statistics of the entire sequence to be tractable during generation. We propose a watermark for masked diffusion language models that controls a global, vector-valued sketch representation of the text. Compared to context-dependent watermarking, the sketch formulation decouples detection from the local contexts seen during generation, resulting in an order-agnostic statistic and a watermarking rule which does not manifest as a simple token bias. We analyze the distortion, soundness, and...
论文介绍 现有研究在机密虚拟机（CVM）中，对文件支持的用户空间可执行对象（如主程序、动态链接库）的运行时完整性保护不足。本文提出了PS-UIE方法，通过特权分离机制来强制执行这些对象的完整性。该方法旨在为CVM提供从启动到运行时的持续完整性保证，解决动态加载执行代码的信任问题，增强了云租户工作负载的安全性。
第一作者: Tianneng Shi · 方向: 软件安全
网络安全AI代理漏洞评估基准测试端到端评估
Abstract:AI has the potential to transform cybersecurity by enabling systems that can autonomously detect, analyze, and remediate software vulnerabilities. However, existing cybersecurity evaluations of AI systems are limited in scale or scope, and fail to capture the end-to-end lifecycle of real-world software vulnerability discovery and remediation. To address this gap, we propose CyberGym-E2E, a large-scale and realistic end-to-end cybersecurity benchmark that comprehensively evaluates AI agents' abilities across the full lifecycle of vulnerability discovery, PoC generation, and patch generation. CyberGym-E2E is comprehensive and scalable, as we build an automated, agent-enhanced pipeline for transforming open-source vulnerability data into realistic evaluation environments. Currently, the benchmark consists of 920 real-world vulnerabilities across 139 different open-source projects.
论文介绍 研究问题是现有AI在网络安全评估中规模和范围有限，无法覆盖真实世界漏洞从发现到修复的完整端到端生命周期。核心方法是提出CyberGym-E2E基准，构建自动化、代理增强的管道，将开源漏洞数据转换为现实评估环境，涵盖920个真实漏洞。意义在于全面评估AI代理在漏洞发现、PoC生成和补丁生成方面的能力，推动网络安全自动化研究。
第一作者: Matthew Finlayson · 方向: 密码学协议
语言模型签名不可伪造性token排序模型安全密码学
Abstract:Language model parameters are known to impose unique (to each model) geometric constraints on their logit outputs, which serves as a signature that identifies the model, but also leaks the model's final layer parameters when an API distributes logits. We investigate more restrictive APIs that expose token rankings (i.e., their ordering by probability, but not the probability values) and find that rankings also constitute a signature: every model has a unique set of feasible top-$k$ rankings for sufficiently large $k$. Furthermore, the ranking signature is the first known (polynomially) unforgeable signature, since finding a model with the same set of feasible rankings is NP-hard. On the security front, we find that token rankings are already sufficient to approximately steal the final layer of the model, similar to logits, though the approximation is too coarse to forge the...
论文介绍 该研究发现，语言模型输出的token排序（仅呈现概率顺序而非具体数值）同样构成一种模型签名。每个模型对于足够大的k，其可行的top-k排序集合是唯一的。该排序签名具有多项式时间内不可伪造的特性，因为寻找一个具有相同可行排序集合的模型是NP困难问题。研究也探讨了token排序可能带来的安全风险，如近似窃取模型最终层参数。
第一作者: José Luis Delgado Jiménez · 方向: 软件安全
密钥封装机制可验证解封装故障识别黑盒测试密码学协议
Abstract:Black-box tests for Fujisaki-Okamoto decapsulation observe the sampled execution seen by the harness, whereas the reencryption computation itself is visible only through the values that reach final key derivation. We study confirmation-code-augmented KEM variants under an honest-reference harness in which the reference encapsulation fixes a hidden final-key point $\langle good,B,W\rangle$, with $W$ the confirmation witness. For a $q$-localized system under test, acceptance is bounded by honest correctness error, adversarial aliasing, final-key freshness defects, a hit on the localized suffix list $Q_G(B)$, and $2^{-\kappa}$. A one-query construction from any predictor of $W$ matches this bound up to the fresh-key coincidence term, so the list-hit event is the black-box obstruction measured by the harness. The list-hit term is bounded either by a cUP-faithful harness...
论文介绍 本文研究了针对基于Fujisaki-Okamoto变换的密钥封装机制的可验证解封装测试。研究分析了在诚实参考测试框架下，测试通过上界的构成因素，包括诚实正确性错误、对抗性混淆、最终密钥新鲜性缺陷等。研究明确了黑盒测试中由列表命中事件构成的主要障碍，并推导了相关的通过界限，为评估相关密码方案的实现安全性提供了理论分析框架。
第一作者: Yuanbo Xie · 方向: AI 安全
智能体系统提示注入跨会话攻击状态持久化LLM安全
Abstract:Modern agentic systems transform LLMs from session-bounded assistants into stateful systems that persist and evolve shared world state across sessions through memories, filesystems, tools, and other long-lived contextual artifacts. This shift fundamentally expands the attack surface of prompt injection. However, prior works on prompt injection have largely focused on model-level threats within a single session, overlooking how cross-session persistent system state fundamentally changes the system-level risk of agentic systems. Inspired by stored cross-site scripting in web systems, we introduce cross-session stored prompt injection, where a successful injection can persist within agentic system state and silently influence future executions long after the original attacker interaction has ended. To systematically study this threat, we formalize stored prompt injection and...
论文介绍 现代智能体系统通过持久化状态跨越会话，这极大扩展了提示注入的攻击面。本文受存储型跨站脚本攻击启发，提出了「跨会话存储提示注入」攻击概念。成功的注入可持久存在于智能体的共享状态中，并在原始攻击交互结束很久后，静默地影响未来的执行。研究形式化了该威胁模型，并探索了其在具有记忆和工具使用能力的智能体系统中的具体风险。
第一作者: Chenghao Li · 方向: 密码学协议
匿名通信广播协议发送者匿名分布式密码学高带宽
Abstract:We present Pepper, a high-bandwidth anonymous broadcast protocol that provides cryptographic sender anonymity against global adversaries. Pepper builds on a two-server DC-net architecture but introduces three key innovations: a self-contained anonymous registration subprotocol using verifiable distributed point functions, support for batch messaging via distributed multi-point functions, and a lightweight access control mechanism based on secret-shared proofs. Unlike prior systems, Pepper eliminates the need for external dialing services and allows each broadcaster to send multiple messages per epoch with a single audit, significantly improving throughput for large data transfers. Our implementation demonstrates that Pepper achieves millisecond-level registration audits, scales efficiently to thousands of channels, and delivers 1.2--20$\times$ higher effective messaging rates...
论文介绍 本文提出了Pepper，一种提供密码学发送者匿名性的高带宽匿名广播协议。它基于双服务器DC-net架构，但创新地引入了使用可验证分布式点函数的自包含匿名注册、支持批量消息发送的分布式多点函数以及基于秘密共享证明的轻量级访问控制。与先前系统相比，Pepper无需外部拨号服务，单次审计可发送多条消息，显著提高了吞吐量，实现了毫秒级的注册审计和可扩展的信道支持。
第一作者: Muhammad Hadi · 方向: 网络安全
联邦学习区块链信任机制资源受限恶意更新过滤
Abstract:Federated Learning (FL) has emerged as an effective paradigm for collaborative intelligence while preserving data privacy. However, data heterogeneity arising from non-IID distributions and decentralized security threats remain significant challenges, particularly in resource-constrained enterprise environments. This paper presents TITAN-FedAnil+, a Trust-Based Adaptive Network for blockchain-enabled federated learning in intelligent enterprises. The proposed framework introduces affinity propagation-based adaptive clustered aggregation to identify and filter malicious updates without requiring prior knowledge of the number of attackers. In addition, GPU-accelerated vectorization is employed to improve computational efficiency, while a signed state jump mechanism enables lightweight blockchain resynchronization. Experimental results demonstrate substantial reductions in memory...
论文介绍 本文针对资源受限企业环境中联邦学习面临的非独立同分布数据异质性与分布式安全威胁问题，提出了TITAN-FedAnil+框架。该框架利用基于亲和传播的自适应聚类聚合来识别和过滤恶意更新，无需预知攻击者数量。同时采用GPU加速向量化提升计算效率，并通过签名状态跳动机制实现轻量级区块链重同步。这些方法旨在提高联邦学习系统的安全性和资源效率。
第一作者: Pritam Dash · 方向: AI 安全
内存中毒LLM智能体攻击分类基准测试AI安全
Abstract:Memory is a core component of AI agents, enabling them to accumulate knowledge across interactions and improve performance. However, persistent memory introduces the risk of memory poisoning, where a single adversarial memory write can exert long-term influence over agent behavior. We present a systematic study of memory poisoning in LLM-based agents. We identify four memory write channels and nine structural vulnerabilities in model capabilities, system prompt design, and agent system architecture that make these channels exploitable. Based on these vulnerabilities, we develop a taxonomy of six classes of memory poisoning attacks. Furthermore, we design MPBench -- a benchmark for evaluating memory poisoning attacks, and show that agents designed to write and retrieve memory more aggressively are more exploitable. We also show that existing prompt injection defenses fail to...
论文介绍 该研究系统性地探讨了基于大语言模型（LLM）的智能体中的内存中毒攻击。论文识别了四个内存写入通道和九个可被利用的结构漏洞，据此提出了一个包含六类攻击的分类法。研究设计了MPBench基准用于评估此类攻击，并发现更激进地读写内存的智能体更容易被攻击。研究还指出，现有的提示注入防御机制无法有效应对此类威胁。
第一作者: Bin Duan · 方向: AI 安全
参数攻击通用防御深度神经网络模型安全重参数化
Abstract:Deep neural networks are increasingly deployed across heterogeneous and partially untrusted environments, where models are distributed through cloud storage, CI/CD pipelines, containerized services, and edge execution platforms. This broad deployment landscape exposes model parameters to various integrity risks. Unlike input-space adversarial attacks, parameter attacks directly tamper with the model's internal parameters and persist across all subsequent inferences. Existing defenses either require retraining, incur significant accuracy degradation, or are limited to specific attack classes. However, in real-world deployment scenarios, the forms of parameter attacks are often unpredictable. To address this challenge, we present ParDef, a generalized defense for deep neural networks against diverse types of parameter attacks. ParDef integrates keyed channel reparameterization...
论文介绍 随着深度神经网络被广泛部署于非受信环境，模型参数面临多样化的完整性风险。现有防御方法通常需要重新训练、精度损失大或仅针对特定攻击类型。本文提出ParDef，一种针对多样化参数攻击的通用防御方法。ParDef通过整合密钥通道重参数化技术，旨在为深度神经网络提供一种不依赖攻击类型预测、且能保持模型精度的鲁棒防护方案。
第一作者: Jeremy Avigad · 方向: 软件安全
形式化验证AIRStark证明Lean 4密码学
Abstract:StarkWare's S-two prover provides an efficient means for establishing, on blockchain, that a program written in the Cairo virtual machine language runs to completion. The latter claim is encoded by an algebraic intermediate representation (AIR) that captures the semantics of the Cairo language. The AIR asserts the existence of tables of values from a finite field satisfying certain algebraic constraints. A cryptographic interactive proof system, circle STARK, provides an efficiently-checked certificate that the AIR is satisfied. We describe our verification, using the Lean 4 proof assistant, that the AIR encoding is sound, which is to say, the satisfiability of the AIR implies the computational claim.
论文介绍 本文描述了对StarkWare的S-two证明器中所使用的代数中间表示（AIR）的形式化验证。AIR编码了Cairo虚拟机语言的运行语义，其可满足性是保障区块链上计算正确性的核心。研究者使用Lean 4证明助手验证了该AIR编码是可靠的，即AIR的可满足性蕴含了底层的计算声明，从而为S-two证明系统的关键组成部分提供了数学上的严格保证。
第一作者: Alireza Sarmadi · 方向: AI 安全
晶体管老化深度神经网络老化感知重训练集成电路可靠性
Abstract:Deep neural networks (DNNs) are used in a variety of real-world applications including, for example, image classification and speech recognition. The inference accuracy of DNN implemented on hardware in integrated circuits (ICs) degrades under phenomena such as transistor aging. Aging slows down the switching speed of transistors, resulting in system-level timing violations due to unsustainable clocks. To maintain reliability for the entire projected lifetime, designers add guardbands to prevent timing violations; however, adding large timing guardbands causes losses in performance (speed or throughput). This chapter provides a detailed discussion of the effects of long-term and short-term transistor aging on DNN inference accuracy. Furthermore, to mitigate aging effects on DNN's accuracy and keep them at bay, a methodology for aging-aware retraining is presented in order to...
论文介绍 本文讨论了集成电路中晶体管的老化现象对部署其上的深度神经网络推理精度的影响。长期和短期的老化会导致晶体管开关速度变慢，引发时序违规。为保障整个生命周期内的可靠性，设计者通常添加时序保护带，但这会牺牲性能。文章详细分析了老化效应，并提出了一种老化感知的重训练方法论，旨在缓解老化对DNN精度的影响，减少对性能保护的依赖。
第一作者: Juan Figuera · 方向: 密码学协议
AI智能体可观测性密码学协议回执信任边界
Abstract:Current AI agent observability is structurally compromised: the entity producing the activity log is the same entity whose activity is being logged. A compromised or buggy agent can omit, alter, or fabricate its own traces, and the operator running the agent has no independent way to detect tampering. We propose a class of protocols that resolves this by inverting the trust boundary: the service that receives an agent's call signs a receipt of what it observed using its own key, encrypts the receipt to the agent's owner, and publishes it to a public transparency log. The owner reconstructs a tamper-evident trail without trusting the agent or its operator. We instantiate the class as Sello, a protocol combining four properties absent in any current system: (P1) receiver-side signing, (P2) HPKE encryption to an owner public key bound to the authorization token via JWS, (P3)...
论文介绍 当前AI智能体的活动日志由同一实体产生和记录，存在被篡改或伪造的风险。本文提出了一类协议来反转信任边界：接收智能体调用的服务方使用自身密钥对观察到的内容签名，生成加密的回执并公开。智能体的所有者因此能在不信任智能体或其运营者的情况下，重构出防篡改的操作轨迹。文中实例化了协议Sello，它结合了接收方签名、加密和公开透明日志等特性。
第一作者: Michael J. Bommarito II · 方向: 软件安全
文件类型分类二进制片段位置无关BERT编码器恶意软件分析
Abstract:File-type classification underlies many workflows like malware triage, forensic carving, packet inspection, and storage indexing. Learned systems such as Google's Magika assume whole-file access at a known offset, so they break on the inputs many of these tasks actually produce, like a single packet payload, a header-less carved fragment, a random disk block, or a chunked upload. We introduce MimeLens, a family of small BERT-style encoders pretrained on binary content from windows sampled at a uniformly random offset within each file, with no privileged head-of-file position, in standard- and short-context variants. A byte chunk goes in from anywhere in a file, no header needed and no fixed size; out comes one of libmagic's 125 MIME labels. On the clean head of complete files, MimeLens beats Magika v1.1 by +10.7 pp top-1 on libmagic-labeled data, and it keeps classifying where...
论文介绍 在恶意软件分析、数字取证等场景中，经常需要对非完整文件或从任意位置开始的二进制片段进行类型分类。现有学习系统（如Magika）通常假设能访问已知偏移的完整文件。本文提出MimeLens，一种基于BERT的小型编码器模型。它通过从文件随机偏移位置采样窗口进行预训练，能够对来自文件任意位置的、无需头部信息的字节块进行内容类型分类，有效解决了片段分类的难题。
第一作者: Kargi Chauhan · 方向: AI 安全
凭据泄露LLM智能体激活探针蜜令牌多轮对话
Abstract:LLM agents often place sensitive credentials in the same context window as untrusted retrieved content, creating a direct path for indirect prompt injection to induce credential exfiltration. We study this failure mode through three complementary defenses. First, we ask whether activation probes can detect credential access before output tokens are emitted. Second, we construct honeytokens from format-specific character models and calibrate detection with split conformal prediction. Third, we treat multi-turn exfiltration as a cumulative information-flow problem and track an estimated leakage budget across conversation turns. In controlled experiments on open-weight models, activation features separate benign and credential-seeking prompts with high accuracy, including under held-out encoding transformations. In a small synthetic multi-turn suite, cumulative accounting detects...
论文介绍 LLM智能体常将敏感凭据与非受信检索内容置于同一上下文窗口，这为间接提示注入攻击诱导凭据泄露创造了路径。本文研究了这一失败模式，并提出三种互补的防御方法：第一，利用激活探针在输出前检测凭据访问；第二，构建格式特定的蜜令牌并使用保形预测校准检测；第三，将多轮泄露视为累积信息流问题，跟踪跨对话轮次的估计泄露预算。这些方法旨在提升智能体的安全性。
第一作者: Avidan Shah · 方向: 安全研究
隐蔽影响语言模型安全负载传播监管失效
Abstract:As language models increasingly consume one another's outputs, covert influence -- a phenomenon where a sender's payload (the behavioral disposition it is conditioned to propagate) transfers to a receiver through carriers undetectable by humans -- becomes a growing risk. We characterize this risk across three interfaces: supervised fine-tuning, on-policy distillation, and in-context learning, and find that they vary in the scale of influence achievable without leaving behind human-visible traces. Using inference-time per-sample attribution scores, we study covert influence across all three interfaces with the ability to select carriers that amplify training-time influence, unlocking payload transfers that prior work could not achieve. We further provide evidence that covert influence with natural-language carriers is a distinct phenomenon from prior studies using number...
论文介绍 本文研究了语言模型之间通过消费彼此输出而产生的「隐蔽影响」风险，即发送方的特定行为倾向（负载）通过人类无法检测的载体传播至接收方。作者系统地评估了在监督微调、策略蒸馏和上下文学习三种接口下的影响规模，并提出利用推理时归因分数来选择能放大影响的载体。研究表明，使用自然语言载体的隐蔽影响是一种有别于过往数值载体研究的独特现象，揭示了当前模型生态中一种新的安全威胁。
第一作者: Sinan Yıldırım · 方向: AI 安全
贝叶斯成员隐私图神经网络隐私成员推理图采样
Abstract:Existing privacy analyses for Graph Neural Networks (GNNs) largely inherit assumptions from non-graph settings, overlooking structural correlations and stochastic training-graph sampling. In particular, node-dependent priors make type-I and type-II errors alone insufficient to characterize the best membership inference test. To address this, we introduce Bayesian Membership Privacy (BMP), a sampling-aware formulation of node-level membership privacy that incorporates node-dependent priors and treats graph sampling probabilities as part of the adversary's knowledge. BMP casts membership inference as a Bayesian hypothesis test and accordingly quantifies membership privacy in terms of posterior membership probability. We explore theoretical properties of BMP in relation to the existing definitions in the literature. We further propose a practical, sampling-aware auditing...
论文介绍 针对图神经网络（GNN）隐私分析中忽视结构相关性和随机图采样的问题，本文提出了「贝叶斯成员隐私」（BMP）这一新的隐私度量框架。BMP将成员推断建模为贝叶斯假设检验，并引入节点相关的先验知识和图采样概率，从而更严谨地量化节点级别的成员隐私。作者进一步提出了一种实用的采样感知审计方法，为在复杂图结构数据上评估GNN的隐私泄露风险提供了更精确的工具。
第一作者: Xinyue Huang · 方向: AI 安全
上下文完整性隐私保护查询重写大语言模型代理任务必要信息
Abstract:As LLMs become increasingly woven into everyday workflows, user queries sent to cloud hosted LLMs routinely mix task-essential content with task non-essential sensitive disclosures, yet type based PII redaction is context agnostic and may raise two issues: over disclosing untyped sensitive context and over removing answer bearing spans. We recast privacy preserving query rewriting under Contextual Integrity: a span should be forwarded only if it is necessary for the task. We introduce DelegateCI-Bench, the first task based Contextual Integrity benchmark for privacy-conscious delegation, comprising 3,167 samples that combine high quality synthetic data spanning 11 tasks and 20 task types, WildChat based real user queries, and a medical challenge set with dense sensitive information. Building on this benchmark, we propose a CI-guided reinforcement learning framework that...
论文介绍 为解决将用户查询发送至云端大语言模型（LLM）时敏感信息过度披露的问题，本文基于「上下文完整性」原则，提出仅转发任务所必需的文本片段。作者构建了首个面向隐私委托任务的CI基准DelegateCI-Bench，并设计了一个由CI引导的强化学习框架来进行查询重写。该方法旨在平衡任务完成度与隐私保护，为实现更安全的LLM工作流代理提供了新的路径。
第一作者: Yingzi Ma · 方向: 安全研究
扩散大语言模型越狱攻击结构自适应攻击安全红队测试
Abstract:Diffusion large language models (dLLMs) generate text by iteratively denoising partially masked sequences under bidirectional context, exposing a safety surface distinct from autoregressive LLMs. Because mask tokens are native inputs and tokens are committed by confidence rather than position, harmful content can be induced through infilling and outside the monitored prefix. Existing jailbreaks either miss this native infill capability or rely on low-diversity mask-bearing templates applied uniformly across goals, with little structural adaptation or accumulated attack experience. We propose MaskForge, a fully black-box adaptive attack that casts dLLM red-teaming as optimized search over a growing library of structural patterns. MaskForge abstracts successful attempts into reusable schemas, selects goal-compatible patterns with a UCB bandit, and invokes a scorer-guided...
论文介绍 本文揭示了扩散型大语言模型（dLLMs）因其迭代去噪机制和掩码输入特性而存在的独特安全风险。作者提出了完全黑盒的自适应攻击方法MaskForge，通过优化搜索一个不断增长的结构模式库来发起越狱攻击。该方法将成功的攻击尝试抽象为可重用模式，并利用UCB赌博机算法选择与目标兼容的模式，从而有效诱导dLLMs生成有害内容，为评估此类模型的安全性提供了新的测试框架。
第一作者: Jack Sanderson · 方向: 软件安全
序列化数据投毒后训练安全多阶段攻击可信AI
Abstract:LLM post-training proceeds through multiple stages, e.g., supervised fine-tuning (SFT) followed by reinforcement learning from human feedback (RLHF) or direct preference optimization (DPO), where each stage draws data from different, potentially untrusted sources. Existing literature assumes data poisoning attacks may occur at each training stage, but neglects the possibility of multiple attackers. To study the trustworthiness of the entire post-training pipeline, we propose the threat model of sequential data poisoning, where multiple adversaries separately poison the SFT and preference datasets. Under this threat model, we identify the single-attacker illusion: each adversary, evaluated in isolation, appears to pose a negligible threat. Yet when adversaries collaborate across stages, the true vulnerability is revealed. In the SFT $\to$ DPO pipeline, their contributions are...
论文介绍 本文研究了大语言模型（LLM）多阶段后训练流程中的安全威胁，提出了「序列化数据投毒」威胁模型。研究发现，单独评估时，分别在监督微调（SFT）和偏好优化（DPO）阶段投毒的单个攻击者看似威胁有限，但两者协同则能暴露系统真正的脆弱性。这挑战了「单攻击者幻觉」，强调了在评估整个后训练管道的可信度时，必须考虑跨阶段的协作攻击场景。
第一作者: Saroj Mishra · 方向: 安全研究
级联幻觉检索增强生成多步推理幻觉检测
Abstract:Multi-step agentic retrieval-augmented generation (RAG) pipelines have demonstrated significant capability for complex reasoning tasks, yet remain vulnerable to a class of failure that existing hallucination detection mechanisms systematically miss: cascading hallucination, where errors introduced at early pipeline stages propagate and amplify across successive reasoning steps, producing confident but factually incorrect final outputs. To address this vulnerability, we formalize cascading hallucination as a distinct failure mode in agentic RAG systems, present a four-type taxonomy of cascade patterns, and introduce CHARM (Cascading Hallucination Aware Resolution and Mitigation), an architectural framework for detecting and interrupting error propagation in multi-step reasoning pipelines. CHARM comprises four components - stage-level fact verification, cross-stage consistency...
论文介绍 针对多步骤代理式检索增强生成（RAG）系统，本文正式定义了现有机制难以检测的「级联幻觉」问题，即早期推理步骤引入的错误会在后续步骤中传播并放大。作者提出了包含四种级联模式的分类法，并介绍了CHARM架构框架，该框架通过阶段级事实验证、跨阶段一致性检查等组件，旨在中断多步推理管线中的错误传播，提升RAG系统的可靠性。
第一作者: Yunsheng Yuan · 方向: 系统安全
隐私保护分布式学习差分隐私非独立同分布数据分布式优化
Abstract:In the paradigm of decentralized learning, a group of agents collaborate to train a global model using distributed datasets without a central server. Although the power of collaboration has been verified by many state-of-the-art studies, it entails extensive gradient information exchanging among the agents and thus induces high risk of privacy leakage for the individual agents. Moreover, in real-world applications, the training data are usually non-identically and independently distributed across the agents, inducing more challenges to enable privacy-preserved decentralized learning. To address these issues, we propose a privacy-preserved decentralized learning algorithm with non-IID data, DPDL, which leverages the notion of Differential Privacy (DP) in cross-gradient aggregation through a similarity-based calibration technique. Specifically, in each round, each agent perturbs...
论文介绍 为解决分布式学习中因梯度交换导致的隐私泄露风险，尤其是在数据非独立同分布（non-IID）的实际场景下，本文提出了DPDL算法。该算法在跨节点梯度聚合过程中引入差分隐私，并采用基于相似性的校准技术来适应non-IID数据分布。DPDL旨在实现隐私保护与模型效用之间的更好平衡，为现实世界中的分布式协作学习提供了更安全的解决方案。
第一作者: Xiaobo Huang · 方向: AI 安全
差分隐私随机梯度下降选择性释放隐私放缩梯度裁剪
Abstract:Machine learning's reliance on sensitive data necessitates privacy-preserving techniques like Differentially Private Stochastic Gradient Descent (DPSGD). However, DPSGD suffers from substantial utility degradation and slow convergence due to gradient clipping and noise injection. Prior works have attempted to improve DPSGD from various perspectives; notably, the Differentially Private Selective Update and Release (DPSUR) algorithm has achieved remarkable model utility. However, the privacy accounting in DPSUR overlooks the variation in sampling probability introduced by the selective release mechanism, which compromises the rigor of its privacy guarantees. To address these limitations, we re-evaluate the privacy analysis of the selective release mechanism and propose a novel algorithm: Differentially Private Selective Release based on Clipped Gradients (DPSR-CG). Through a...
论文介绍 差分隐私随机梯度下降（DPSGD）及其变体DPSUR存在隐私分析不够严格的问题。本文重新审视了选择性释放机制中的隐私放缩效应，并指出DPSUR忽略了选择性释放引入的采样概率变化。为此，作者提出了DPSR-CG算法，通过更精确的隐私分析来修正这一缺陷，旨在在保护严格隐私保证的前提下，寻求对模型效用影响最小的改进方案。
第一作者: Xixi Tian · 方向: 系统安全
联邦学习脓毒症早期预测隐私保护多中心医疗数据
Abstract:Privacy-sensitive and distributed characteristics of multi-center medical data bring severe obstacles to centralized modeling for accurate early prediction of sepsis. Federated learning (FL) has attracted growing attention as a promising framework for collaborative model development, as it allows multiple institutions to jointly train predictive models without directly sharing or centralizing raw data. Nevertheless, its practical performance, robustness, and privacy-preserving benefits remain insufficiently evaluated using real-world clinical datasets. To bridge this gap, this study systematically examines the application of federated learning to multi-center sepsis prediction. The experimental dataset consists of 648 clinically screened samples collected from three tertiary hospitals in China, with rigorous inclusion and exclusion criteria. We establish a centralized training...
论文介绍 本研究探讨了联邦学习在多中心脓毒症早期预测中的应用。研究问题在于医疗数据隐私敏感且分散，难以进行集中建模。核心方法是采用联邦学习框架，允许三家中国三级医院在不共享原始数据的情况下协同训练预测模型，并利用临床筛选的样本进行系统评估。该研究旨在验证联邦学习在真实临床场景下的性能、鲁棒性和隐私保护效果。
第一作者: Bochen Lyu · 方向: 安全研究
大语言模型安全对齐自回归一致性攻击
Abstract:Safety alignment in large language models (LLMs) is fragile in part because it is often shallow: fine-tuning mainly reshapes the model's behavior near the first few output tokens. We argue that this phenomenon can be understood through autoregressive consistency, the tendency of next-token prediction to preserve and extend the current response trajectory consistently. By analyzing the learning dynamics of safety alignment, we show that autoregressive consistency can concentrate alignment updates on early tokens, offering a mechanistic explanation for shallow safety alignment. The same mechanism also predicts a broader class of attacks on LLMs: attacks that induce harmful continuation states at arbitrary positions in the output trajectory. As a concrete example, we introduce random insertion attack, which inserts a short harmful span into an otherwise safe refusal trajectory...
论文介绍 该研究指出了大语言模型安全对齐的脆弱性及其机制。核心观点认为，安全对齐之所以浅层，是因为自回归生成特性会将模型更新集中在输出序列的初始部分。基于此机制，论文预测并提出了一类新的攻击方式，例如随机插入攻击，可在输出轨迹的任意位置诱导有害续写，从而绕过安全防护。研究为理解与加固LLM安全提供了新视角。
第一作者: Zexun Wang · 方向: 安全研究
代理系统运行时治理动作证书模型无关
Abstract:Agent systems execute through runtimes with very different control points: local coding tools, framework SDKs, managed agent platforms, API gateways, and observer-only integrations. A high-risk action such as publishing data externally may therefore appear as a shell command in one runtime, a tool call in another, and a hosted session transition in a third. This makes it difficult to answer a basic governance question consistently: what action was authorized, under whose authority, with what approval semantics, and with what evidence after execution? This paper presents Proof-Carrying Agent Actions (PCAA), a runtime-neutral governance model centered on an action certificate rather than on a vendor-native session record. PCAA organizes control around five checkpoints: pre-action admissibility, action open, assumption capture, approval, and outcome closure. It binds these...
论文介绍 针对异构代理系统中高风险动作治理不一致的问题，本文提出了“携带证明的代理动作”模型。该模型以中立于特定运行时的“动作证书”为核心，而非厂商原生的会话记录，从而统一回答动作授权、权威、审批语义及执行证据等关键治理问题。它通过五个检查点组织控制流程，旨在为多样化的代理执行环境提供一致的运行时治理框架。
第一作者: Wei Liu · 方向: AI 安全
大语言模型强化学习奖励函数社会规则规则漏洞
Abstract:Reinforcement learning (RL) has become a dominant post-training paradigm, enabling large language models (LLMs) to learn from rewards. We observe that societal regulations are structurally similar to reward functions. They define measurable outcomes, thresholds, and exceptions, while often leaving institutional intent only partially specified. We hypothesise that the RL training process may exploit these gaps and therefore ask whether models' well-known tendency to hack reward functions during RL can scale into a more consequential failure mode named societal hacking: discovering loopholes in the rules society runs on. To study this phenomenon, we introduce SocioHack, a sandbox of 72 societal environments, and find that within these environments, reward hacking naturally emerges and leads to regulatory loophole discovery. Models learn to hack the social rules and generate...
论文介绍 本研究假设，强化学习中模型利用奖励函数漏洞的“黑客行为”，可能扩展到对社会规则的“黑客行为”。论文将社会法规类比为奖励函数，认为RL训练过程可能利用规则中的未明确部分。为此，作者构建了包含72个社会环境的沙盒SocioHack，实验发现模型能在其中自动发现监管漏洞并生成规避方案，这揭示了AI安全的一个潜在新风险维度。
第一作者: Xian Qi Loye · 方向: AI 安全
代理安全基于评分标准的强化学习工具使用安全细粒度奖励
Abstract:The evolution of LLMs into tool-enabled agents creates a new class of safety challenges associated with real-world execution rather than simple text generation. Existing alignment methods often rely on coarse refusal signals or static supervision, making it difficult to balance safety with useful tool execution across diverse agentic risks. We introduce RUBAS, a rubric-based reinforcement learning framework for agent safety. RUBAS decomposes agent behavior into four dimensions: tool-use safety, argument safety, response safety, and helpfulness. These structured rubrics provide fine-grained and interpretable rewards over complete agent trajectories, enabling reinforcement learning to optimize safe tool use while preserving task completion. Extensive experiments across multiple agent safety benchmarks and models show that RUBAS improves safety over standard alignment baselines...
论文介绍 本文针对工具增强型LLM代理的安全挑战，提出了RUBAS框架。核心方法是将代理行为分解为工具使用安全、参数安全、响应安全和有帮助性四个维度，并基于评分标准提供细粒度的奖励信号，用于强化学习优化。这旨在平衡安全与任务完成能力，相比依赖粗糙拒绝信号的传统对齐方法，RUBAS提供了更精细、可解释的安全训练途径。
第一作者: Tianyi Xie · 方向: 机器人操作 · 来源: cs.RO
人形机器人移动操作数字生成流程视频先验
Abstract:Scaling humanoid loco-manipulation requires robot-compatible demonstrations across diverse objects, whole-body motions, and scene geometries, but teleoperation and motion capture are difficult to scale because each collection depends on physical setups, instrumented actors, and robot operation. We present GRAIL, a digital generation pipeline that remains fully virtual until deployment: it composes 3D assets, simulator-ready scenes, and priors from video foundation models (VFMs) to synthesize interactions without rebuilding physical environments or teleoperating the robot. Rather than reconstructing unconstrained in-the-wild videos, GRAIL starts from fully specified 3D configurations in which object geometry, camera parameters, metric scale, environment depth, and a robot-proportioned character are known before video generation and reused during reconstruction. This privileged...
论文介绍 为解决人形机器人移动操作数据收集成本高的问题，本文提出了GRAIL数字生成流程。核心方法是在完全虚拟的环境中，组合3D资产、仿真场景和视频基础模型的先验知识来合成交互数据，无需物理环境重建或机器人遥操作。该方法从已知的3D配置出发生成视频并用于重建，旨在高效生成机器人兼容的演示数据，以扩展技能学习的规模。
第一作者: Rachel Luo · 方向: 数据集与评测 · 来源: cs.RO
策略评估神经替代模型方差减少异构数据
Abstract:Rigorous evaluation of learning-based robotic systems is an essential prerequisite for deployment. However, real-world test data is expensive to gather; moreover, in a typical iterative development context, data gathered from the latest policy is necessarily limited in scale. This motivates evaluation methodologies that make use of heterogeneous data sources, including simulation, historical policy logs, and data collected from related platforms or environments. While such auxiliary data are abundant and inexpensive, they are generally not directly representative of real-world outcomes -- for example, performance in simulation may differ substantially from performance in the real world -- making their principled use for high-confidence performance estimation challenging. In this paper, we introduce X4Val, a general framework for variance-reduced real-world metric estimation in...
论文介绍 在机器人系统评估中，真实世界测试数据昂贵且有限。本文提出X4Val框架，旨在利用仿真、历史数据等廉价但非真实的辅助数据，进行高置信度的性能估计。核心方法是学习神经网络替代模型，通过减少估计方差来提升评估精度。该框架为在数据受限情况下，有效利用异构数据源进行严谨的系统评估提供了通用解决方案。
第一作者: Chenhao Bai · 方向: 导航与运动 · 来源: cs.RO
课程学习物理域扩展可恢复性四足运动
Abstract:Scaling robust robot policies requires more than broader randomization, because physical-domain experience must remain organized and learnable throughout training. We study when a policy can benefit from harder physics and identify recoverability as a central constraint in on-policy physical-domain scaling. In on-policy training, new dynamics are useful only insofar as they remain close enough to the current policy to generate corrective on-policy data, rather than collapsing rollouts into unrecoverable failures. Using quadruped locomotion as a physically demanding benchmark for embodied generalization, we introduce HORIZON, a checkpointed frontier curriculum that expands physical domains only within the current policy's recoverable boundary. HORIZON uses rollback and boundary refinement to govern each expansion step, turning fixed randomization into a continual process of...
论文介绍 本研究指出，扩展机器人的物理域随机化训练时，需关注新动态的可恢复性。核心贡献是提出HORIZON课程框架，它只将物理域扩展到当前策略能生成可纠正数据、避免不可恢复失败的范围内。该方法使用回滚和边界精炼来管理每一步扩展，将固定的随机化转变为持续的过程。实验以四足运动为基准，展示了该课程在稳定学习复杂动力学方面的有效性。
第一作者: Luca Zanatta · 方向: 导航与运动 · 来源: cs.RO
世界模型环境变异性视觉导航四旋翼无人机强化学习
Abstract:World models, learned generative models that predict how an environment evolves, have become a promising tool for sample-efficient robot learning. Yet how robust they are to environmental variability remains poorly understood. To address this, we conduct a systematic study using vision-based quadrotor navigation as a testbed problem, training DreamerV3-based world models under varying levels of environmental randomness and evaluating them across all levels through cross-environment validation, spanning both Self-Supervised Learning (SSL) pretraining and Reinforcement Learning (RL) fine-tuning. We then deploy all world models and associated navigation policies on a real quadrotor in unseen environments, including an open-loop run where the model receives just 2.5s of real sensory input before all sensors are cut off, leaving the system to navigate entirely in imagination over a...
论文介绍 本文研究世界模型在环境变异性下的泛化能力，以视觉四旋翼导航为测试平台。采用基于DreamerV3的世界模型，在不同环境随机性下训练并通过跨环境验证评估，结合自监督学习预训练和强化学习微调。实验在真实四旋翼上部署，展示在未见环境中的导航性能，包括开环模式下的想象导航。研究有助于提升机器人学习在复杂环境中的鲁棒性。
第一作者: Yunpeng Mei · 方向: VLA 通用模型 · 来源: cs.RO
视觉-语言-动作流匹配策略改进自引导成功率轨迹
Abstract:Large vision-language-action (VLA) policies are increasingly trained as conditional generative models over action chunks. Yet deployment produces mixed-quality experience-successful demonstrations, partial completions, recoverable mistakes, and failures-that is difficult to use with standard imitation. Full behavior cloning (BC) imitates failures, filtered BC discards useful sub-trajectories, and offline reinforcement learning adds a large critic. We introduce ForesightFlow, a self-guided flow-matching policy that augments each generated action chunk with a learned success-potential trajectory. The same flow proposes and scores candidate actions, enabling best-of-$K$ inference without an external critic. The key issue is that policy improvement and value calibration require different supervision: advantage weighting should emphasize high-quality actions, but applying the same...
论文介绍 本文针对大型视觉-语言-动作策略在部署中产生混合质量经验的问题，提出ForesightFlow方法。该方法采用自引导流匹配策略，为每个动作块生成成功率轨迹，实现无需外部评论家的best-of-K推理。通过统一的流提出和评分候选动作，解决策略改进与价值校准的不同监督需求，提升模仿学习的样本效率和策略性能。
第一作者: Ning Yang · 方向: 导航与运动 · 来源: cs.RO
视觉导航世界-动作模型潜空间扩散变换器具身导航
Abstract:Visual navigation requires generating smooth and collision-free trajectories under complex geometric and physical constraints. Existing reactive policies that directly map observations to actions lack anticipatory reasoning, limiting their ability to proactively avoid obstacles. While visual imagination offers predictive foresight, conventional modular approaches separate scene prediction from policy learning, often leading to error accumulation and inefficient inference. To address these limitations, we propose WAM-Nav, a Latent World-Action Model for embodied visual navigation that jointly learns action generation and latent visual foresight, enabling more robust and foresighted navigation decisions without compromising inference efficiency. Specifically, WAM-Nav utilizes a shared Diffusion Transformer for asymmetric joint diffusion to concurrently generate long-horizon...
论文介绍 本文提出WAM-Nav，一种用于具身视觉导航的潜空间世界-动作模型。该模型联合学习动作生成和潜空间视觉预见，通过共享扩散变换器进行非对称联合扩散，生成长时程轨迹。相比传统模块化方法，WAM-Nav避免误差累积，提高推理效率，实现更鲁棒和预见性的导航决策。
第一作者: Abhishek S · 方向: 导航与运动 · 来源: cs.RO
不确定性感知SLAM可靠性估计LiDARRGB-D相机
Abstract:Low-cost 2D LiDARs lack the intensity channel that higher-end sensors use to diagnose measurement failures, yet they are widely used on educational and budget robotics platforms. We present SENTINEL, a training - free, label - free reliability estimation framework that gives range - only LiDAR an effective diagnostic signal. SENTINEL combines geometry-based scan statistics with cross - modal depth consistency between LiDAR and an RGB - D camera to compute a per - scan reliability score between 0 and 1. When the score falls below a threshold, corrupted scans are rejected and the robot falls back to calibrated wheel odometry, preventing silent SLAM corruption. We evaluate SENTINEL on a GEFIER R1 four - wheel skid-steer robot equipped with an RPLidar A2M12 and an Intel RealSense D435i in a 185 cm by 245 cm arena containing controlled transparent and reflective failure elements on...
论文介绍 本文针对低成本2D LiDAR在SLAM中易受测量失败影响的问题，提出SENTINEL可靠性估计框架。该框架无需训练或标签，结合几何扫描统计和跨模态深度一致性，为每次扫描计算可靠性分数。当检测到不可靠扫描时，系统拒绝该数据并回退到轮式里程计，防止SLAM无声损坏。实验在真实机器人平台上验证，提升导航系统的安全性。
第一作者: Zuxing Lu · 方向: 机器人操作 · 来源: cs.RO
全身控制多模态模仿人形机器人运动模仿共享潜空间
Abstract:Building a general-purpose whole-body controller is essential for enabling diverse motion capabilities in humanoid robots across a wide range of downstream tasks, including locomotion and loco-manipulation. Different tasks rely on distinct motion reference modalities: locomotion primarily depends on coordinated robot joint trajectories, whereas manipulation requires precise end-effector trajectory tracking. Existing methods often overlook the representational mismatch between dense robot joint angles and sparse end-effector poses. To address this, we propose Multi-Modal Mimic (M3imic), a versatile multi-modal whole-body control framework that unifies heterogeneous motion reference modalities, including robot joint angles, human pose trajectories, and end-effector poses, using modality-specific encoders to map them into a shared latent space. Leveraging large-scale...
论文介绍 本文提出M3imic，一个用于人形机器人的多模态全身控制框架。该框架统一异构运动参考模态，包括关节角度、人类姿态轨迹和末端执行器姿态，通过模态特定编码器映射到共享潜空间。解决表示不匹配问题，实现从大规模数据中学习通用控制器，支持移动和操作等多样化任务。
第一作者: Amirhosein Alian · 方向: VLA 通用模型 · 来源: cs.RO
触觉感知视觉-触觉数据集模仿学习触觉反馈机器人操作
Abstract:Despite the importance of tactile sensing for reliable manipulation, most existing Vision-Language-Action (VLA) datasets remain vision-only, and those that do incorporate tactile information typically lack the joint combination of task diversity, language conditioning, and action trajectories. Furthermore, existing teleoperation pipelines rarely provide haptic feedback to the operator, despite its established role in demonstration quality and manipulation stability. In this work, we present HapTile, a contact-grounded visuotactile manipulation dataset that advances beyond vision-only trajectory datasets by embedding physical interaction sensing at two levels: fingertip tactile feedback at the robot end-effector, and haptic-informed demonstrations at the teleoperator side. The data collection platform integrates haptic feedback directly into the teleoperation controller...
论文介绍 本文介绍HapTile，一个用于接触丰富模仿学习的视觉-触觉-语言-动作数据集。该数据集超越纯视觉数据集，在机器人末端执行器和遥操作端嵌入触觉反馈，提供任务多样性、语言条件和动作轨迹的联合组合。数据收集平台集成触觉反馈，改善演示质量和操作稳定性，推动机器人操作研究。
第一作者: Achilleas Santi Seisa · 方向: 具身智能 · 来源: cs.RO
5G连接边缘控制空中机器人工业矿井模型预测控制
Abstract:This article presents the first real-world autonomous flight of a 5G-connected aerial robot controlled by an edge-offloaded controller, and aims to bridge the gap between controlled and factual setups. The robot operates within an active industrial subterranean mine, while the high-level controller is deployed in a nearby Kubernetes-based edge cluster. Communication between the robot and the edge is enabled via a 5G New Radio (NR) Standalone (SA) network. The chosen controller is a Model Predictive Controller (MPC), which generates control actions to allow the robot to navigate seamlessly through the mining environment. A human operator selects waypoints for the aerial robot, and the MPC generates smooth, collision-free paths for autonomous executions. The proposed 5G edge-based closed-loop system is evaluated in a real industrial setting and demonstrates the potential of...
论文介绍 本文首次在工业地下矿井中部署5G连接的空中机器人，实现边缘控制下的自主飞行。机器人通过5G NR SA网络与附近边缘集群通信，使用模型预测控制生成平滑无碰撞路径。人类操作员选择航点，系统自主执行。实验在真实工业环境中评估，展示5G边缘控制在复杂场景中的应用潜力。
第一作者: Guopeng Li · 方向: 策略学习 · 来源: cs.RO
安全强化学习Q学习Cholesky投影样本效率机器人控制
Abstract:Safe robot control requires maximizing return while satisfying safety constraints. In off-policy safe reinforcement learning, reward and safety Q-values are commonly learned by separate critic ensembles, with uncertainty handled independently for each objective. This objective-wise treatment neglects inter-objective correlation and can lead to overly conservative value estimates, thereby reducing sample efficiency. To address this issue, we propose Cholesky-Ordered Projection Q-learning (COP-Q), a safety-first method that incorporates inter-objective covariance into vector-valued Q-value estimation. COP-Q constructs a generalized confidence bound in the joint Q-value space and uses Cholesky factorization to encode objective priority in a sequential form. This preserves conservatism on safety while adaptively reducing excessive conservatism on the reward objective. The...
论文介绍 本文提出COP-Q方法，用于安全强化学习中的机器人控制。该方法在离线设置中，通过向量值Q值估计整合目标间协方差，使用Cholesky投影构造联合Q值空间的广义置信界。COP-Q保持安全约束的保守性，同时自适应减少奖励目标的过度保守性，提升样本效率，实现更高效的安全控制策略学习。
第一作者: Abhishek S · 方向: 数据集与评测 · 来源: cs.RO
多智能体路径规划执行时间预测评价指标协调结构CADENCE
Abstract:Multi-Agent Path Finding (MAPF) algorithms are increasingly used to plan motion for robot teams in industrial warehouses and robotic shared workspaces, but standard MAPF algorithm evaluation metrics, such as Sum of Costs (SoC), makespan, and planner runtime, can obscure how planner choices translate into realistic execution performance. We present CADENCE (Coordination and Action-Driven Estimation for Networked Continuous Execution), a hardware study of this evaluation gap on a fixed 7 by 7 workcell with seven differential drive robots, asking which features available before execution can best predict final wall-clock completion time. We compare SoC, total planned travel cost, primitive motion burden (how much basic motion the plan requires, such as makespan, turns, consecutive moves, and start-stop transitions), and interaction aware coordination structure (how much...
论文介绍 多智能体路径规划算法在工业仓库等场景广泛应用，但标准评估指标如总成本（SoC）和Makespan难以反映真实执行性能。本文提出CADENCE框架，通过硬件实验在7x7工作单元中使用七个差速驱动机器人，分析执行前特征如基本运动负担和协调结构，以预测最终完成时间，旨在优化机器人团队任务规划的评价体系。
第一作者: Kailun Huang · 方向: 导航与运动 · 来源: cs.RO
类人机器人步态适应多地形运动混合专家模型强化学习
Abstract:Humans primarily rely on walking and running to traverse complex terrains, without resorting to unnecessarily complex motion patterns. Similarly, humanoid robots should achieve smooth transitions between walking and running while maintaining natural and stable locomotion. However, unifying gait transition and multi-terrain adaptation within a single policy remains challenging due to gradient interference and the distribution shift induced by terrain-dependent visual and dynamic variations. Although Mixture-of-Experts (MoE) architectures can alleviate multi-skill interference, naive joint training often fails to yield clear expert specialization, limiting their effectiveness. To address these challenges, we propose CoRe-MoE, a two-stage reinforcement learning framework that decouples gait generation from terrain adaptation. In the first stage, a stable locomotion policy is...
论文介绍 类人机器人在复杂地形中需实现行走和奔跑的平滑过渡并保持运动稳定性，但统一步态转换和地形适应策略因梯度干扰和分布偏移而困难。本文提出CoRe-MoE，一种两阶段强化学习框架，将步态生成与地形适应解耦，通过对比重加权混合专家模型提高专家特化能力，以实现多地形下的自然运动。
第一作者: Siyuan Yang · 方向: VLA 通用模型 · 来源: cs.RO
通用操作接口视觉语言动作模型数据适应物理验证VISTA
Abstract:Universal Manipulation Interface (UMI) enables scalable real-world robot data collection without hardware-specific teleoperation, yet leveraging UMI data to train large-scale Vision-Language-Action (VLA) models remains fundamentally challenging. We identify two critical mismatches: wrist-mounted fisheye views, with severe radial distortion and local gripper-centric perspectives, are out-of-distribution for pretrained VLMs; and human-collected trajectories frequently violate kinematic limits, incur collisions, or exceed controller bandwidth, teaching VLA policies physically infeasible actions. To address the challenges, we present VISTA, a framework that bridges this dual gap through three synergistic components. (i)~UMI-VQA, the first large-scale VQA dataset tailored to wrist-mounted fisheye observations, aligns VLM representations to the distorted visual regime via auxiliary...
论文介绍 通用操作接口（UMI）支持可扩展的真实机器人数据收集，但用于训练视觉语言动作（VLA）模型时存在挑战：腕部鱼眼视图失真且局部视角偏离预训练VLM分布，人类收集轨迹常违反运动学限制。本文提出VISTA框架，通过UMI-VQA数据集对齐视觉表征、物理验证确保动作可行性等组件，提升VLA模型在真实世界中的泛化能力。
第一作者: Abhishek S · 方向: 导航与运动 · 来源: cs.RO
地下矿井自主导航强化学习开源基准模拟环境
Abstract:Underground mines present extreme conditions for autonomous robot navigation: GPS is denied, lighting is degraded, and tunnel topology is loop-rich and non-convex. Simulation benchmarks grounded in real production-mine geometry and compatible with GPU-accelerated learning pipelines do not yet exist in the open-source ecosystem. We present MineXplore, an open-source MuJoCo-based navigation benchmark derived from the Leung et al. 2017 Chilean underground copper mine dataset. The environment reconstructs a 104,423 sq.m tunnel network through an six-stage contour-to-MJCF pipeline incorporating octagonal wall cross-sections, LiDAR-sourced jagged wall geometry, three terrain friction zones, a global 5 degree incline, and periodic spot lighting. Geometric fidelity is validated at an Intersection over Union (IoU) of 0.9538 against the source survey map, and surface texture similarity...
论文介绍 地下矿井环境对自主机器人导航构成挑战，如GPS失效、照明不足和复杂隧道拓扑，目前缺乏基于真实矿井几何的开源模拟基准。本文提出MineXplore，一个基于MuJoCo的开源导航基准，重建104,423平方米隧道网络，包含八边形墙壁截面等特征，并支持GPU加速学习，以促进GNSS拒止环境下的探索研究。
第一作者: Zhuoyuan Wu · 方向: 策略学习 · 来源: cs.RO
世界模型动作条件跨机器人泛化策略评估数据标准化
Abstract:We present OSCAR, a precise action-conditioned video world model that generalizes across different robot embodiments and enables robot policy evaluation. Existing video world models face three main challenges for real-world robot evaluation: limited scenario diversity in current robot training datasets, imprecise action following, and poor generalization across embodiments for broad adoption. We tackle these challenges from two perspectives. At its core is a large-scale standardized data pipeline that curates, filters, and deduplicates broad robotics and egocentric human datasets, yielding a clean joint-training dataset that spans diverse tasks, scenarios, actions, and robot embodiments. To condition the video model, we adopt 2D kinematic skeleton rendering as a unified conditioning representation that generalizes across different robot arms or even human hands. We finetune...
论文介绍 现有视频世界模型在机器人策略评估中面临场景多样性有限、动作跟踪不精确和跨机器人泛化差的挑战。本文提出OSCAR，一个精确的动作条件视频世界模型，通过大规模标准化数据管道整合多样机器人数据集，并使用2D运动学骨架渲染作为统一条件表示，实现跨不同机器人 embodiment 的泛化，以支持策略评估和视频生成。
第一作者: Yilong Wang · 方向: 机器人操作 · 来源: cs.RO
可变形物体操作上下文模仿学习流匹配变换器零样本迁移视觉表示
Abstract:Deformable object manipulation (DOM) is challenging due to high-dimensional, partially observable states that evolve through long-horizon, topology-changing interactions with multiple valid manipulation modes. We introduce Instant-Fold, an in-context imitation learning framework for DOM. Given a single human demonstration, our policy infers and executes diverse manipulation modes directly from the demonstration, including variations in spatial execution and ordering, without requiring gradient updates. Our approach first learns deformation-aware visual representations via temporal contrastive pretraining, after which a flow-matching transformer policy conditioned on the demonstration predicts actions to execute the intended manipulation mode. Trained entirely in simulation, Instant-Fold generalizes across diverse folding modes and transfers zero-shot to real-world settings...
论文介绍 可变形物体操作因状态高维、部分可观测和拓扑变化而具有挑战性。本文提出Instant-Fold，一种上下文模仿学习框架。给定单个人类演示，策略能直接推断并执行多种操作模式，无需梯度更新，通过时间对比预学习变形感知视觉表示和流匹配变换器预测动作，在模拟中训练后零样本迁移到真实世界。
第一作者: Tianchong Jiang · 方向: 机器人操作 · 来源: cs.RO
机器人操作基准评估缺陷失败模式诊断方法统计显著性
Abstract:A robotics benchmark score measures success under one fixed evaluation setup, yet is routinely treated as evidence of general manipulation capability. We identify four failure modes, each of which weakens or invalidates a benchmark's role as a valid proxy for that capability: shortcut solvability, lack of statistical significance, creeping overfitting, and data-source dependence. We propose one diagnostic per failure mode. We audit LIBERO, CALVIN, SimplerEnv, RoboCasa, and RoboTwin 2.0 under these diagnostics. LIBERO and CALVIN fail multiple diagnostics. RoboCasa and RoboTwin 2.0 fail fewer, despite appearing far less often in recent progress claims. On LIBERO, a 0.09B probe with no language encoder scores at or near reported SOTA, and most reported gains are not provably statistically significant. On CALVIN, randomizing block poses within the training range drops performance...
论文介绍 机器人操作基准分数常被用作通用能力的证据，但可能无效。本文识别出四种失败模式：捷径可解性、缺乏统计显著性、渐进过拟合和数据源依赖。针对每种模式提出诊断方法，并审计LIBERO、CALVIN等基准，发现LIBERO和CALVIN存在多个问题，强调基准评估需更严格验证以准确衡量操作能力。
第一作者: Charlie Gauthier · 方向: 策略学习 · 来源: cs.RO
语义场景重建大语言模型规划模拟环境自动构建计划验证
Abstract:Simulation environments are useful for both robot policy learning and planning verification and validation. Traditionally, the process of creating a simulation was onerous. Creating a bespoke simulation environment for each individual environment that a robot would operate in was simply infeasible. In this work, we introduce PerceptTwin, a fully automatic pipeline that constructs interactive simulations directly from semantic scene representations produced by a robot's perception stack. PerceptTwin combines open-vocabulary object maps with 3D asset generation, affordance prediction, and commonsense condition checking. These interactive simulations can be used to validate and refine plans before they are executed on the robot hardware. Borrowing from the AI alignment literature, we also introduce an LLM judge that verifies plan correctness and alignment with human preferences...
论文介绍 模拟环境对机器人策略学习和计划验证至关重要，但传统方法创建过程繁琐。本文提出PerceptTwin，一个全自动管道，直接从机器人感知栈的语义场景表示构建交互式模拟，结合开放词汇对象地图、3D资产生成等技术，用于在硬件执行前验证和精炼计划，并引入LLM裁判验证计划正确性。
第一作者: Junyi Cao · 方向: 机器人操作 · 来源: cs.RO
可变形线性物体机器人操作可微物理模拟基准测试
Abstract:We address the challenge of enabling robots to manipulate deformable linear objects (DLOs), such as ropes, cables, and rubber bands. Prior work has primarily focused on narrow, task-specific problems, often relying on real-world demonstrations or handcrafted heuristics. Such approaches, however, struggle to scale to the wide variety of materials and tasks encountered in practice, and collecting sufficiently diverse real-world data is often impractical. Additionally, existing simulation environments offer limited support for the broad spectrum of material behaviors necessary for generalizable DLO manipulation. To overcome these limitations, we introduce a differentiable simulator explicitly designed for versatile DLO manipulation. Our simulator models a wide range of material properties-including (in)extensibility, elasticity, bending plasticity, and complex interactions with...
论文介绍 针对机器人操作可变形线性物体（如绳子）这一难题，现有方法常局限于特定任务，且缺乏多样化的模拟环境。本文提出一个专为此类操作设计的可微物理模拟器，能建模广泛的材料特性（如弹性、塑性弯曲）及复杂交互，旨在支持通用化研究与基准评估。
第一作者: Amirhossein Mollaei Khass · 方向: 导航与运动 · 来源: cs.RO
多智能体风险规避规划视角优化分布式协调高斯溅射
Abstract:Multi-agent Next-Best-View (NBV) selection for safe path planning in uncertain and unknown environments requires informative, safety-aware, and efficient coordination. Centralized approaches rely on sharing raw sensor data or significant communication overhead, resulting in limited scalability. We propose a distributed, risk-aware multi-agent NBV framework in which each robot maintains a private local 3D Gaussian Splatting map and the team jointly maximizes expected information gain (EIG) restricted to masked zones along planned trajectories. The resulting distributed objective is solved by Consensus ADMM (C-ADMM) over a communication graph, with each robot exchanging only candidate viewpoints, planned trajectory descriptors, and scalar EIG contributions. Collision risk along each trajectory is modeled via Average Value-at-Risk (AV@R) over the local 3DGS map and used both to...
论文介绍 在未知环境中进行多智能体路径规划需要兼顾信息增益与安全性。本文提出一个分布式、风险感知的多智能体“下一最优视角”框架。每个机器人维护私有3D高斯溅射地图，并通过共识ADMM算法协调，在规划轨迹上最大化受限于遮罩区域的信息增益，同时使用平均风险值来评估并规避碰撞风险。
第一作者: Tewodros Ayalew · 方向: 模仿学习 · 来源: cs.RO
世界模型潜在动作自监督学习模仿学习视频生成
Abstract:We introduce CLAW, a fully end-to-end self-supervised framework for learning a world model jointly with continuous latent action representations directly from action-free videos. Our approach leverages adversarial latent regularization and diffusion-based video generation to capture structured and semantically meaningful action representations while modeling rich, predictive environment dynamics, without relying on any action labels or annotations. By simultaneously training the Latent Action Model and world model, CLAW learns to reason about how inferred actions induce environment transitions from visual observations alone. We show that the resulting latent action world model supports both imitation learning from observation and goal-directed planning. In imitation learning, latent actions extracted from raw videos enable behavior cloning. For planning, CLAW generates...
论文介绍 本文提出CLAW框架，能够直接从无动作标签的视频中，通过对抗性潜在正则化和基于扩散的视频生成，联合学习世界模型与连续的潜在动作表示。该模型可从视觉观测推理动作如何引发环境变化，并支持从观测中进行行为克隆的模仿学习以及目标导向的规划。
第一作者: Jiaxin Shi · 方向: VLA 通用模型 · 来源: cs.RO
视觉-语言-动作模型3D空间推理3D几何感知联合训练
Abstract:We propose a 3D-thinking-guided co-training framework that enables vision-language-action (VLA) models to perform 3D spatial reasoning implicitly during action prediction. Our core insight is that 3D geometry perception and 3D spatial reasoning are distinct capabilities that can be disentangled and injected at different feature hierarchies. During training, three tightly coupled components work in concert primarily within the latent space: (1) To gain geometric priors, a latent 3D geometry perception module aligns intermediate visual features with a 3D foundation model, acquiring low-level geometric cues without architectural modifications to the VLM backbone. (2) Complementing this, an online 3D reasoning distillation module mitigates the prompt-induced reasoning gap via a shared reasoning anchor token. During 3D VLM co-training, this anchor is emitted as the first output...
论文介绍 为提升视觉-语言-动作（VLA）模型的3D空间推理能力，本文提出一种3D思维引导的联合训练框架。其核心在于将3D几何感知与空间推理能力解耦并注入模型：通过潜在3D几何感知模块获取低级几何线索，并利用在线3D推理蒸馏模块来缓解推理差距，从而在动作预测过程中实现隐式的3D空间推理。
第一作者: Boyuan Xiao · 方向: VLA 通用模型 · 来源: cs.RO
视觉语言决策感知瓶颈焦点计划场景理解机器人操作
Abstract:In embodied vision-language decision making tasks such as robotic manipulation and navigation, Vision-Language and Vision-Language-Action Models (VLMs & VLAs) are powerful tools with different benefits: VLMs are better at long-term planning, while VLAs are better at reactive control. However, their performance is limited by the same perceptual bottleneck: visual hallucinations arise due to the models' inability to distinguish task-relevant objects from distractors. In principle, accurate identification and focus on critical objects while filtering out irrelevant ones is the key to break this limitation. A straightforward solution is one-step focus: directly attending to essential objects. However, this approach proves ineffective because effective focus inherently requires deep scene understanding. To this end, we propose SceneDiver, a coarse-to-fine focus plan generation...
论文介绍 在机器人操作等视觉语言决策任务中，模型因无法有效区分关键物体与干扰物而产生视觉幻觉，性能受限。本文提出SceneDiver方法，通过生成一个从粗到细的“焦点计划”来引导模型进行深度场景理解，从而突破这一感知瓶颈，帮助模型准确识别并专注于任务关键对象。
第一作者: Yu Zhu · 方向: 策略学习 · 来源: cs.CV
食品分析视觉语言模型强化学习多任务学习卡路里推理
Abstract:Recent studies have explored Vision-Language Models (VLMs) for food analysis. However, most existing methods rely primarily on supervised fine-tuning (SFT), which often limits reasoning and generalization capabilities. Moreover, high-quality large-scale nutritional annotations remain scarce. To address these issues, we introduce CalorieBench-80K, a large-scale benchmark with curated calorie labels and dietary advice annotations. To the best of our knowledge, it is the first food image benchmark to incorporate Chain-of-Thought (CoT) annotations for calorie reasoning. We also propose Food-R1, a unified food VLM trained in a multi-task learning paradigm to equip the model with broad capabilities. Food-R1 undergoes CoT-based cold-start instruction tuning, followed by reinforcement fine-tuning (RFT) using Group Relative Policy Optimization (GRPO) to improve reasoning and...
论文介绍 针对现有食品视觉语言模型依赖监督微调、推理能力有限的问题，本文提出Food-R1。该模型基于新构建的、包含思维链注释的CalorieBench-80K基准，在多任务学习范式下训练，并通过基于强化学习的微调（使用GRPO算法）来增强模型的推理与泛化能力，用于统一的食品分析任务。
第一作者: Rui Zhao · 方向: 机器人操作 · 来源: cs.CV
视频生成机器人操作物理接地可执行动作
Abstract:Video generation models have made impressive strides in synthesizing visually compelling content, yet their outputs remain confined to the virtual domain. A natural question follows: how well do these models reflect the physical world when their generated videos leave the screen and enter reality? We propose robotic manipulation as a concrete, measurable window onto this question: if a model has truly internalized physical laws, the motion it depicts should translate into executable robot behavior. We introduce Dream$.$exe, an evaluation framework that operationalizes this criterion through a video-to-execution pipeline. Given a scene image and a task description, Dream$.$exe synthesizes a manipulation video, converts the generated motion into robot trajectories, and executes them in a physics simulator, yielding a grounding signal that purely visual metrics cannot offer...
论文介绍 视频生成模型能合成逼真内容，但其输出是否真实反映物理世界规律尚不明确。本文提出一个评估框架Dream.exe，以机器人操作作为检验窗口：给定场景图像和任务描述，生成操作视频，将其中描绘的运动转化为机器人轨迹，并在物理模拟器中执行，从而为生成模型提供纯粹视觉指标无法给出的物理接地信号。
第一作者: Mutian Tong · 方向: 机器人操作 · 来源: cs.CV
3D点表示视频-动作模型机器人控制4D建模
Abstract:Video-Action Models (VAMs) leverage the broad visual dynamics captured by pre-trained video diffusion models, offering a promising path toward generalizable robot manipulation. However, RGB-only video rollouts are not directly actionable: they leave metric 3D motion, contact geometry, and fine-grained spatial constraints under-specified, making action grounding ambiguous. Meanwhile, scaling action supervision across diverse tasks and embodiments remains costly. We present PointAction, a framework that bridges video predictions to robot actions through explicit point-based 4D modeling. PointAction fine-tunes a foundation video generation model to jointly predict future RGB frames and dynamic 3D pointmaps, producing temporally consistent 3D motion of task-relevant scene geometry. These point dynamics serve as a structured, embodiment-agnostic action interface, which a...
论文介绍 基于RGB视频预测的机器人操作在动作接地方面存在模糊性。本文提出PointAction框架，通过显式的基于点的4D建模来连接视频预测与机器人动作。该框架微调基础视频生成模型，使其能联合预测未来RGB帧与动态3D点图，产生时序一致的场景点动态。这些点动态作为结构化、与具身体无关的动作接口，进而用于生成机器人动作。
第一作者: Xinhong Zhang · 方向: 策略学习 · 来源: cs.AI
可微分模拟GPU加速四旋翼控制策略学习并行仿真
Abstract:This letter introduces DiffAero, a lightweight, GPU-accelerated, and fully differentiable simulation framework designed for efficient quadrotor control policy learning. DiffAero supports both environment-level and agent-level parallelism and integrates multiple dynamics models, customizable sensor stacks (IMU, depth camera, and LiDAR), and diverse flight tasks within a unified, GPU-native training interface. By fully parallelizing both physics and rendering on the GPU, DiffAero eliminates CPU-GPU data transfer bottlenecks and delivers orders-of-magnitude improvements in simulation throughput. In contrast to existing simulators, DiffAero not only provides high-performance simulation but also serves as a research platform for exploring differentiable and hybrid learning algorithms. Extensive benchmarks and real-world flight experiments demonstrate that DiffAero and hybrid...
论文介绍 该论文介绍了 DiffAero，一个轻量级、GPU 加速且完全可微分的模拟框架，旨在解决四旋翼控制策略学习中的效率瓶颈。该框架支持环境级和代理级并行性，集成多种动力学模型与可定制传感器，通过完全并行化物理和渲染消除 CPU-GPU 数据传输限制，显著提升模拟吞吐量。DiffAero 不仅为高性能模拟提供平台，还可用于探索可微分和混合学习算法，在四旋翼自主飞行等任务中具有应用潜力。
美股方面，SPY和QQQ显示bullish趋势，SPY接近52周高点且多头排列，但QQQ的RSI 73.5超买与MACD死叉提示短期压力；部分大型科技股如MSFT、META出现MACD金叉但趋势中性。加密市场极度恐慌，恐慌贪婪指数为12，BTC、ETH、SOL的RSI均低于20处于超卖，价格远低于所有均线呈空头排列，总市值2.27万亿美元中BTC主导率55.8%显示资金集中。中概股如BABA、PDD趋势bearish，RSI偏低但未超卖，0700.HK短期反弹但长期结构仍弱。商品外汇中，黄金和原油期货中性，美元指数DXY呈bullish多头排列接近52周高，美元/人民币接近52周低。整体市场分化，美股相对强势，加密超卖严重，中概偏弱，商品外汇混合。
当前价格63169.83远低于SMA20（74015.36）、SMA50（76650.91）和SMA200（78929.52）形成空头排列。RSI14为17.6处于超卖状态，MACD值-3004.34且信号线-1616.95显示空头动量持续。近5日跌幅14.35%，接近52周低点仅高5.15%，技术面显著偏弱。
价格757.09接近52周高点仅低0.44%，RSI14为69.7未进入超买。SMA20（745.99）、SMA50（711.9）和SMA200（683.44）呈多头排列，近1日涨0.38%显示短期动量。MACD值11.95虽低于信号线12.58，但趋势定义为bullish。
RSI14为73.5进入超买区域，MACD出现死叉（21.06 vs 21.51）。但价格740.61仍高于SMA20（721.5）、SMA50（665.46）和SMA200（621.14），趋势为bullish，多头排列信号仍在。超买与死叉提示短期回调风险，与长期趋势形成矛盾。
MACD出现金叉（7.11 vs 6.35），显示动量转向积极。但趋势定义为neutral，RSI14为52.9处于中性区域。价格428.05在SMA20（422.79）附近，低于SMA200（456.85），技术面无明显方向性突破。
MACD出现金叉（-7.85 vs -12.01）指示动量改善，近5日涨幅8%显著。但价格459仍低于SMA20（453.15）、SMA50（478.03）和SMA200（572.62）为空头排列，趋势bearish。RSI14为49.2中性，短期反弹与长期下行趋势并存。
VIX 恐慌指数
10Y 美债收益率 (%)
美元指数 DXY
S&P 500 ETF
Nasdaq 100 ETF
Apple
Microsoft
Nvidia
Alphabet
Tesla
Meta
Bitcoin
Ethereum
Solana
阿里巴巴 (BABA)
拼多多 (PDD)
京东 (JD)
腾讯控股 (0700.HK)
黄金期货
WTI 原油期货
美元 / 人民币
本报告基于公开历史技术指标数据生成，过去走势不代表未来表现，仅供技术指标解读参考，不构成任何投资建议。技术分析存在固有局限性，实际交易需结合基本面和个人风险承受能力。
The insect, which has been found in America for the first time since 1966, is a growing concern in the southern US.
中文摘要 螺旋蝇幼虫自1966年后首次在美国被发现，美国计划使用不育蝇和猎犬来对抗这一在美国南部日益令人担忧的疫情。
Legislation would also sanction key segments of Russian economy, overriding objections from Republican leaders The House passed legislation Thursday that would aid Ukraine and sanction key segments of the Russian economy, overriding objections from Republican leaders who warned the bill would underm
中文摘要 美国众议院通过一项向乌克兰提供援助并对俄罗斯经济关键领域实施制裁的法案，此举无视了共和党领导人的反对，是一周内第二次与特朗普立场相左。
Follow today’s news live Get our breaking news email, free app or daily news podcast Jane Hume, the deputy opposition leader, accused the government of being “sneaky”, saying discussions around tax reforms in parliament were much too short to properly address so-called “generational” tax reforms. “W
中文摘要 澳大利亚反对党副领袖指责政府在税务改革讨论中行事“鬼祟”。同时，前总理霍华德签署公开信敦促进行赌博改革。
The New York Times has been investigating the number of boat strikes and the death toll in the Caribbean and Eastern Pacific carried out by the Trump administration. Eric Schmitt, a national security correspondent for The Times, explains what those numbers mean.
中文摘要 《纽约时报》调查特朗普政府在加勒比海和东太平洋进行的船只打击次数及死亡人数。该报国家安全记者解释这些数字的含义。
Hezbollah’s leader said the Iran-backed group, which was not included in U.S.-brokered talks, said a truce worked out between Israel and the Lebanese government amounted to surrender for his group.
中文摘要 真主党拒绝美国斡旋下以色列与黎巴嫩政府达成的停火协议，称该协议对其组织而言等于投降。
As concern rises in Europe over threats from an emboldened Russia, the Irish government says it’s working to plug gaps in its military, which reflect a tradition of neutrality.
中文摘要 随着欧洲对俄罗斯威胁的担忧加剧，爱尔兰政府表示正努力加强其军队，以弥补因中立传统而形成的防务缺口。
Police allege the 44-year-old lauded for disarming one of the gunmen in December’s attack put his father in a headlock Follow our Australia news live blog for latest updates Get our breaking news email, free app or daily news podcast The Sydney man who disarmed one of the Bondi terror attackers has
中文摘要 曾因在悉尼邦迪袭击中制服一名枪手而受到称赞的艾哈迈德·艾哈迈德，因涉嫌对父亲实施锁头控制而被指控袭击，他对此表示“震惊”。
The invertebrates were likely destined for the pet trade, including as reptile food Follow our Australia news live blog for latest updates Get our breaking news email, free app or daily news podcast More than 100,000 live exotic cockroaches have been seized from a commercial breeder in New South Wal
中文摘要 澳大利亚新南威尔士州破获一起重大案件，从一家商业饲养场查获超过10万只活体异国蟑螂，价值约20万美元，这些蟑螂可能用于宠物贸易。
Eighteen G.O.P. lawmakers broke with their party and joined Democrats to deliver yet another blow to the president’s foreign policy agenda.
中文摘要 美国众议院不顾共和党领导人反对，通过了乌克兰援助法案。18名共和党议员与民主党人一同投票，再次打击了总统的外交政策议程。
Israel's supreme court annuls government ban on international Red Cross visiting Palestinian prisoners in Israeli jails.
中文摘要 以色列最高法院撤销了政府禁止国际红十字会探视以色列监狱中巴勒斯坦囚犯的禁令。同时，真主党拒绝停火协议，以色列继续对黎巴嫩发动袭击。
BBC Verify examines how the biggest change to the White House in decades has transformed in the last year.
中文摘要 BBC核实报道，特朗普政府提出的白宫舞厅改造计划在过去一年里，其规模和成本均已翻倍。
“After 26 years in power, age is beginning to take its toll,” President Volodymyr Zelensky of Ukraine wrote of his Russian counterpart, bragging of a recent strike on St. Petersburg.
中文摘要 乌克兰总统泽连斯基在致俄罗斯总统普京的信中混合了嘲讽与和谈提议，他提及普京执政26年，年龄开始产生影响，并夸耀了对圣彼得堡的打击。
Ukraine's president tells the Russian leader that only "direct engagement" between the two countries could end the war, with the US focused on Iran.
中文摘要 乌克兰总统泽连斯基在一封公开信中向俄罗斯总统普京提议进行面对面会谈，称两国之间的“直接接触”是结束战争的唯一途径。
Congressman Ro Khanna says the integration push only boosts Netanyahu, but his push to revoke the measure fell short.
中文摘要 美国民主党议员罗·卡纳试图撤销加深美以军事合作的提案失败。他认为，此举只会增强以色列总理内塔尼亚胡的影响力。
Swampy mangrove forests are staging a surprise comeback - which is good news for coastal communities and the climate.
中文摘要 BBC报道，经过数十年的人类破坏，湿地红树林正在意外地恢复，这对沿海社区和气候而言都是好消息。
Broadcom CEO Hock Tan discusses the "leap of faith" the company took when starting its work with Anthropic on chips a year ago, describing it as a "great bet." (Source: Bloomberg)
中文摘要 博通首席执行官陈福阳表示，公司一年前与Anthropic在芯片领域的合作是一个巨大的赌注，是一次在人工智能领域的“信仰之跃”。
Blistering gains in Vietnam’s biggest stock have confounded seasoned analysts and set off a frenzy among retail traders, raising doubts about how long the rally can last.
中文摘要 越南最大上市公司Vingroup的股价飙升1000%，引发散户交易狂潮并让资深分析师感到困惑，市场质疑此轮涨势的可持续性。
The rally in China’s semiconductor stocks is poised to extend thanks to upcoming blockbuster IPOs and technology breakthroughs, investors and analysts say.
中文摘要 投资者与分析师认为，中国半导体股票的9000亿美元市值上涨行情将得以延续，动力来自即将进行的巨额IPO和技术突破。
The president is directing money into coal as the Iran war drives up energy costs for Americans.
Global investors are rapidly losing confidence in Indonesia as the nation’s stocks tumble at the fastest pace worldwide and its currency sinks to all-time lows.
中文摘要 随着印尼股市跌幅和货币贬值创下全球最快纪录，“卖出印尼”的情绪在交易台蔓延，全球投资者对该国信心迅速流失。
The overarching narrative is that Indian equity markets missed out on the global artificial intelligence boom. But a look under the hood reveals a slew of smaller firms winning from trillions of dollars being spent on AI capacity.
中文摘要 一只不知名的印度股票年内大涨530%，揭示了在全球AI投资热潮下，印度一些中小公司正在成为隐藏的赢家。
Gold was little changed as uncertainty persisted over progress in US-Iran talks to end the war that’s roiled global markets.
中文摘要 由于美国与伊朗为结束战争而进行的谈判进展存在不确定性，金价在动荡的市场环境中保持稳定。
Instead of throwing away byproducts of food processing, fermentation is making them valuable.
中文摘要 发酵等古老技术正被用于将食品加工副产物转化为有价值的产品，以减少食物浪费。
Jackie Jantos says loneliness and lack of confidence were challenges for young adults looking for relationships.
中文摘要 约会应用Hinge负责人表示，20多岁的单身人士在恋爱初期面临孤独和信心不足的挑战，需要AI帮助迈出第一步。
Naomi Gleit has weathered many controversies at Meta, but remains in what she tells the BBC is her "dream job".
中文摘要 Meta任职时间最长的员工Naomi Gleit回顾了她在公司经历的多次争议，并称这是她的“梦想工作”。
Sempra Infrastructure said it began producing liquefied natural gas from Mexico’s first West Coast export terminal, with plans to ultimately ship the fuel to Asia in a move that could add relief to a market hurt by tightened global supply due to the Iran war.
中文摘要 Sempra基础设施公司开始在墨西哥首个西海岸出口终端生产液化天然气，计划未来将其运往亚洲，这可能缓解因伊朗战争导致的全球供应紧张。
Mira Murati, Co-Founder & CEO of Thinking Machines Lab, discusses the future of human-AI interaction, evolving AI products and the biggest opportunities ahead with Bloomberg’s Emily Chang at Bloomberg Tech 2026 in San Francisco. (Source: Bloomberg)
中文摘要 Thinking Machines联合创始人兼首席执行官Mira Murati在Bloomberg Tech 2026上讨论了人机交互的未来、AI产品的演进以及最大的机遇。
4 回复 · 程序员 节点
7 回复 · 程序员 节点
15 回复 · 程序员 节点
9 回复 · 程序员 节点
5 回复 · 程序员 节点
7 回复 · 程序员 节点
11 回复 · Apple 节点
10 回复 · Apple 节点
26 回复 · Apple 节点
17 回复 · Apple 节点
今天升到2级同时又是我生日，双喜临门啊 不过过了半个小时，为什么系统还没判定今天登录过? 90 个帖子 - 89 位参与者 阅读完整话题
应该没问题了 muyuan.do 不要在我这里浪费时间，去做你该做的事情，去爱你该爱的人 158 个帖子 - 156 位参与者 阅读完整话题
L站的佬友们好，我入行这个AI漫剧这行已经三个月了，基本都是接触的海外项目，然后这个月刚做了海外阅文和抖音合作的海外项目，在审核阶段，就和大家分享一下自己做真人海外的心得 人物站位和人物走位控制 情绪参考 工具使用 画面，镜头的参考 最后，大家做这个真人时候，我建议大家看着提示词过一遍画面，脑子里面有这个画面的详细情况，你能知道这个画面是不是符合你的想法，还有就是，剪辑，配乐真的很重要，有能力的佬友可以去抖音找找suno ai音乐的教程，我这几天都在做这个的广告项目 欢迎佬友来问我问题，但是我上班时间一般不会登陆L站，但是晚上9点钟后到家一般会回复，还有如果是小白要入这行现在，我的建议是，先按
l0veyou公益站推出了数字人模型，一次可生成10秒，效果非常逼真，没过多久我会给它支持到生成20秒一个视频，这样更好 在这里选择音频（上传的音频需要大于两秒，否则会生成视频失败），然后点击上传模型，我需要在右下角的加号，点击上传参考图，然后提示词可以这样填：她在说话。 另外提醒一下， Ai生图暂时用不了了。我今天修了一天也没修好，可能至少要明天才能好了。今晚也不知道还有没有得睡，因为明天我还要上架新模型，明天我要上架gemini3.1pro（0.1ldc一句话）和3d模型生成（生成出来的3d模型，它的零件可以直接拆分，而不是一整块模型，效果非常好） 14 个帖子 - 9 位参与者 阅读完整
感情是相爱的通行货币，无法估值，无法退款。在一起时能够兑换幸福，分别之时，将瞬间破产，成了欠款，利息即为眼泪偿还。如果她在就好，回来就好，不愿这生活暗无天日。 55 个帖子 - 37 位参与者 阅读完整话题
停机时间延长至今晚10点 161 个帖子 - 154 位参与者 阅读完整话题
今日早上八点半奥特曼重置了所有账号，当前正在管理9个Pro20x，刚好可以拿来估计一波GPT订阅的额度。 账号： 到截图时间的用量数据： 模型 输入 输出 缓存读取 合计 Token GPT5.5 62,178,157 1,759,623 245,985,792 309,923,572 GPT5.4 284,049,752 12,110,928 1,555,673,600 1,851,834,280 平均缓存命中率： 83.2% 到截图时间的费用数据： GPT5.4：1313 美元 到截图时间9个账号合计使用周限制： 7% + 9% + 8% + 8% + 8% + 7% + 11% + 10
转自新华社 记者从教育部获悉，今年全国高考报名人数为1290万人。 对于高校招生宣传工作，教育部强调，严禁虚假宣传、违规承诺，严禁以任何形式炒作“高考状元”“高分考生”“升学率”。推动高校录取通知书回归“一页纸”，坚决纠治奢华录取通知书、新生礼盒等不良风气。 来源：新华社微信公众号综合人民日报客户端（记者：吴丹） 有佬友知道这是何意味吗？ 86 个帖子 - 38 位参与者 阅读完整话题
考虑休息一会儿 16 个帖子 - 15 位参与者 阅读完整话题
我现在是浑身冒火 44 个帖子 - 42 位参与者 阅读完整话题