独立 AI 资讯库
从自动抓取数据库实时读取，适合快速浏览、检索和后续扩展成完整资讯站。
arXiv:2607.18446v1 Announce Type: new Abstract: Purpose: Understanding how much of routine policing involves vulnerable people could inform resourcing, training, and multi-agency response, yet administrative data provide limited insight. We explore whether an LLM-based classification pipeline, deve
arXiv:2607.18443v1 Announce Type: new Abstract: Pragmatic language use requires reasoning about alternatives: the alternative expressions a speaker might have chosen, or the alternative interpretations a listener might entertain. Formal and computational models of pragmatics must therefore specify
arXiv:2607.18438v1 Announce Type: new Abstract: Introducing Relay-Bench, an unsaturated, holistic, text-only benchmark that measures LLMs' ability to complete an assortment of tasks from distinct domains in a single prompt. The leading model, GPT-5.5 (xHigh), scores 43.3%. The test set entirely con
arXiv:2607.18432v1 Announce Type: new Abstract: This paper reports on a collaboration between the Directorate-General for Translation (DGT) and the European Master's in Translation (EMT) to localise the MMLU dataset into 11 European languages. Beyond creating a more inclusive benchmark for LLM eval
arXiv:2607.18366v1 Announce Type: new Abstract: Large language models (LLMs) serving as planners in tool-using autonomous agents introduce dynamic reliability risks in multi-turn execution. While single-turn safety mechanisms are relatively mature, extended interactions reveal structural vulnerabil
arXiv:2607.18270v1 Announce Type: new Abstract: While Electronic Health Records (EHRs) offer a wealth of clinical data, effectively augmenting a patient's records with heterogeneous external knowledge to predict the patient's clinical risk remains a significant challenge. Existing methods fail to c
arXiv:2607.18266v1 Announce Type: new Abstract: Small language models are attractive for local deployment, but they often struggle with multi-step arithmetic reasoning. We study whether structured synthetic reasoning data can improve this behaviour under consumer-hardware constraints. Starting from
arXiv:2607.18265v1 Announce Type: new Abstract: Long-running Large Language Model (LLM)-based agents often accumulate large intermediate traces containing audits, eliminations, and numeric calculations. In practice, this state is compressed before handing it to a downstream decision step, creating
arXiv:2607.18264v1 Announce Type: new Abstract: Language models solve complex problems by articulating intermediate reasoning steps in natural language. While effective, this process is computationally bottlenecked: each reasoning step conveys only a single subword, and many are spent expressing a
arXiv:2607.18261v1 Announce Type: new Abstract: LLM agents are increasingly used as transaction compilers: a user states an intent in natural language, and the model emits a structured object that an API can execute. JSON Schema and provider-level structured-output modes are useful because they rem
arXiv:2607.18260v1 Announce Type: new Abstract: We introduce FindStatBench, an execution benchmark for evaluating large language models on combinatorial code synthesis. Built from FindStat, it contains 2,329 tasks across 24 collections and 5.52M hidden instances, covering statistic synthesis, which
arXiv:2607.18255v1 Announce Type: new Abstract: Contribution attribution has become a central problem in LLM-based multi-agent systems, where final outputs are produced through multiple agents, message exchanges, and ordered workflow dependencies. Existing attribution methods often rely on counterf
arXiv:2607.18246v1 Announce Type: new Abstract: We present Phionyx, a deterministic AI runtime architecture derived from the broader Echoism interaction framework that introduces a governance-first approach to AI engineering: treating large language model (LLM) outputs as noisy sensor measurements
arXiv:2607.18245v1 Announce Type: new Abstract: Exact-match evaluation of agent-calling obscures qualitatively different failure modes: a model may select the right function yet hallucinate argument values, or satisfy a schema while choosing a agent for the wrong reason. Existing benchmarks collaps
arXiv:2607.18243v1 Announce Type: new Abstract: Agentic AI is crossing trust boundaries faster than current risk models can represent. Existing approaches provide one of two partial views. They either describe failure mechanisms without producing a transferable residual-risk estimate, or they produ
arXiv:2607.18242v1 Announce Type: new Abstract: The coming era of autonomous AI agents demands a discovery mechanism capable of navigating millions of tools, yet existing solutions buckle under O(N) complexity and centralized governance. Instead of building another fragile overlay, we propose ToolD
arXiv:2607.18239v1 Announce Type: new Abstract: Power-seeking defined as behaviors where AI systems acquire resources, evade oversight, or resist termination beyond task requirements is identified as a key driver of Loss of Control (LoC) risk. In this work, we introduce SysAdmin, a benchmark that p
本文介绍了如何借助 AWS PrivateLink 与 Direct Connect/VPN 将本地企业数据源安全私密地连接至 Amazon Quick，彻底避免分析与交互流量暴露于公网。架构将访问链路划分为面向用户的终端访问流量与面向数据源的数据拉取流量，全面覆盖 Quick Sight、Quick Agents、MCP 服务及自动化流程等组件。对于受 HIPAA、FedRAMP、PCI-DSS 等严苛合规要求约束或追求低延迟、高审计能力的企业，该方案提供了兼顾 AI 洞察与数据网络边界安全的实操指南。
构建多 Agent 系统的原型如今仅需数天，但要将其推进到真实生产环境中，开发者往往会在编排、模型选型、运行可靠性、安全代码执行和监控运维这五大方面遭遇瓶颈。针对这一痛点，AWS 解决方案架构团队开源了生产级多 Agent 系统 Deep Insight，全面展示了如何落地 Anthropic 提出的上下文工程技术。该方案借助 Strands Agents SDK 的 Graph 与 Agents-as-Tools 架构、基于 Amazon Bedrock 为不同角色的 Agent 按需匹配 Claude 模型，并通过 AWS Fargate 沙箱与 Bedrock AgentCore 实现安全代码隔离和托管运维。该系统为开发者跨越 AI 原型与企业级落地的鸿沟提供了兼顾性能、成本与安全性的可复用工程参考范式。
2026 上海开源软件应用创新大赛「智算云赛道」决赛入围名单今日公布，以下 10 支队伍成功晋级，将于 10 月 29 日在上海张江人工智能创新小镇参加总决赛。 现将入围名单按报名时间顺序公示如下： 项目名称 团队负责人 代码仓库地址 ShannonBase 李浩 https://github.com/Shannon-Data/ShannonBase 女娲 Nuwax AIOS 邱超 ...
当前构建 AI Agent 的企业普遍面临单次请求成本高、延迟不确定以及无法离线运行的挑战。基于 Strands Agents 构建的自适应智能体架构（AAIF）提出了一种分级处理新方案：利用端侧运行的小语言模型（SLM）处理日常状态检查与视觉分析，仅将复杂的跨领域分析和多步推理路由至 Amazon Bedrock 云端模型。实践表明，该架构能将约 80% 的常规流量截留在本地，不仅显著降低了 API 调用成本并实现低延迟，还支持在树莓派等边缘设备、混合云及纯离线网络间灵活适配，为工业级 Agent 系统的工程化落地提供了重要参考。
OpenAI 发布声明，阐述其如何与分发合作伙伴协作以维持一致的安全保障措施。官方通过明确《使用政策》（Usage Policies）中的预期要求，确保在不同分发渠道下均能实现负责任的 AI 模型访问。此举旨在为庞大的生态渠道树立统一的安全底线与合规标准，为下游开发者和企业客户跨平台调用 OpenAI 模型提供了清晰的合规预期。
Hackers Employ Grok and Claude to Attack LatAm Banks PYMNTS.com
10 月 30 日，以“从开源创新到全球增长”为主题的 2026 上海开源软件应用创新大赛颁奖典礼暨创新增长大会，将落地上海张江人工智能创新小镇。 2026上海开源软件应用创新大赛由开源中国主办，以“开源向实（Open Source, Real Impact）”为主题，面向全国征集优秀开源项目。自大赛启动报名以来，累计已收到报名项目 700 ...
在主要竞争对手亚马逊封锁 AI Agent 代替用户购物之际，沃尔玛反其道而行之，正式宣布推出 Agent Gateway 及一系列配套技术以拥抱智能体商务（Agentic Commerce）。该网关基于 Anthropic 开创的模型上下文协议（MCP）构建为 MCP 服务器，支持获得批准的 AI Agent 在保护用户隐私的前提下检索库存并代客下单。此外，沃尔玛还通过“Agent Profiles”功能为用户提供类似信用卡副卡的额度控制、交易逐笔审批与误购退货保障。这一开放策略不仅与封禁 Agent 的主流电商平台形成鲜明对比，也为 ChatGPT、Gemini 等 AI 开发者和生态伙伴打通了实体零售交易的关键闭环。
Grok 4.7 Is Now in Viktor: What You Need to Know BASENOR
大模型训练过程，开始更多地出现在公众视野中。此前，小米MiMo负责人罗福莉公开了MiMo-V2.6强化学习阶段的训练进展，让外界看到了大模型训练中的一些具体细节。现在，Physical AI领域也有了类似尝试。10月21日，生数科技世界模型团队正式开启Motubrain 2.0训练全流程公开直播，计划持续半个月，向公众展示具身模型的训练过程及研究进展。这次直播关注一个问题：Physical AI究竟如何实现？Physical AI，即物理人工智能，希望让AI能够理解物理环境，并通过机器人完成实际任务。相比大语言模型，Physical AI 还需要解决动作决策、任务执行和跨场景泛化等问题。Mot
美国国家安全局（NSA）紧急预警，PHP 中曝出两个可追溯至 2004 年（PHP 5）的严重漏洞，波及全球约 70% 的网站（超 3300 万个站点）。该漏洞由安全团队借助 OpenAI GPT-5.6 Sol 以及 Wiz Atlas 安全 AI Agent、Gemini 4 Argon 等工具联合发现并协同修复，允许攻击者渗透服务器文件系统或数据库。虽然此前因利用难度极高未出现野外攻击，但大模型大幅降低了自动化生成漏洞利用代码的门槛，NSA 建议无法即刻修补的网站应直接下线。这标志着 AI 正在以双刃剑的姿态重塑网络安全生态——既能提前消除深层隐患，也因其武器化潜力放大了未修补资产的风险。
SpaceX’s AI Surprise Is More Fragile Than It Looks Bloomberg.com
AT&T 正在与 OpenAI 展开合作，致力于让企业法务知识更易于获取与检索。其 LegalEdge 团队正重新构想法务需求受理流程，并借助大语言模型规模化扩展法务人员的专业判断力。该实践展示了大型电信企业如何将生成式 AI 深度融入严谨的法务合规场景，为企业内部法律 Agent 与智能化工作流提供了参考借鉴。