🤖 AI 速览

OpenAI 一边推出面向关键基础设施的 Daybreak 计划，强化网络防御与行业落地，一边用 GPT-6 Astra 进入法律检索、财务审阅和游戏原型场景。今天的重点不在模型噱头，而在 AI 正加速嵌入专业流程，工程瓶颈也更清晰地暴露在评测、记忆和可信执行上。
📋 文章元数据
发布时间
2026-09-04
类型
ai-daily
字数
5968
阅读时长
29 min

2026-09-04 AI日更 | OpenAI 以 10 亿美元押注 AI 安全防御，Astra 进入法律与游戏工作流 链接到标题

OpenAI 一边推出面向关键基础设施的 Daybreak 计划，强化网络防御与行业落地，一边用 GPT-6 Astra 进入法律检索、财务审阅和游戏原型场景。今天的重点不在模型噱头，而在 AI 正加速嵌入专业流程，工程瓶颈也更清晰地暴露在评测、记忆和可信执行上。

📖 本期 Watch List 深度导读 链接到标题

今天最值得跟进的，是“AI 正在从通用模型走向可落地系统”这一条主线。Google DeepMind 的 WeatherNext 3 代表基础模型继续向高精度行业预测推进，值得看它如何把天气这种强时空问题做成可部署能力；OpenAI 则把 GPT-6 Astra 用到法律检索和游戏原型里，说明 agent 正在进入专业工作流和生产效率场景。另一组论文更值得工程团队细读：围绕评测觉察、持久记忆、静态 LLM API 的数据层、以及多 agent 证据可信度的研究，集中暴露了当前智能体系统最现实的工程瓶颈。

🌐 X 平台 AI 热点快讯 链接到标题

话题 1:OpenAI Launches GPT-6 Astra as Most Capable Model Yet 链接到标题

  • 分类:AI · News
  • 概况:热度时间:5 hours ago,相关帖子数:79000
  • 是什么事:X 上热议 OpenAI 预发布名为 Astra 的新模型安全更新，并传出其可能在近日上线、能力显著强于 GPT-5.6 的消息。
  • 为什么重要:如果属实，这意味着 OpenAI 可能在模型能力、推理效率和网络安全攻防能力上再次拉开差距，也会影响行业对新一代前沿模型发布时间、命名和发布节奏的判断。
  • 讨论概况:讨论焦点主要集中在 Astra 是否已进入最终发布阶段、真实公开名称到底是 GPT-6 还是其他版本、以及泄露信息和官方信息之间有多大可信度；另一部分争议在于现有公开证据更多支持其在网络安全场景突破，而非通用能力全面跃升。

话题 2:Lululemon Shares Plunge 15% After Weak Earnings and Cut Outlook 链接到标题

  • 分类:AI · News
  • 概况:热度时间:3 hours ago,相关帖子数:3800
  • 是什么事:Lululemon 发布低于预期的财报并下调业绩展望，股价盘中大跌约 15%。
  • 为什么重要:这类消费品牌的业绩与指引变化，常被视为宏观需求、零售数据和市场风险偏好的风向标，也会影响 AI 相关投资对消费科技与零售自动化叙事的估值判断。
  • 讨论概况:X 上的讨论主要集中在业绩走弱的原因是需求放缓、竞争加剧还是库存与折扣压力，也有人争论这只是短期波动，还是公司增长逻辑已经明显降温。

话题 3:JD Vance Blames Iran Attacks for High Gas Prices in White House Briefing 链接到标题

  • 分类:AI · Other
  • 概况:热度时间:1 day ago,相关帖子数:37000
  • 是什么事:白宫简报会上，JD Vance 将高油价归因于对伊朗的袭击，引发 X 平台上的广泛关注与转述。
  • 为什么重要:这类表态会影响市场对能源价格、地缘政治风险和政策叙事的判断，而这些变量也会间接影响 AI 产业的算力成本、供应链预期和宏观投资环境。
  • 讨论概况:X 上的讨论主要集中在这番归因是否成立、是否是在为油价上涨寻找政治解释，以及伊朗局势与能源市场之间的实际关联有多大。

话题 4:John Ternus Takes Over as Apple’s New CEO from Tim Cook 链接到标题

  • 分类:AI · News
  • 概况:热度时间:2 days ago,相关帖子数:267000
  • 是什么事:有消息称，John Ternus 接替 Tim Cook 出任苹果新任 CEO。
  • 为什么重要:苹果是全球最重要的科技公司之一，其管理层变化会直接影响 AI 产品路线、芯片策略、设备生态和行业竞争格局。
  • 讨论概况:X 上的讨论主要集中在这是否意味着苹果将更激进推进端侧 AI、是否会调整现有的保守产品节奏，以及新 CEO 能否在 AI 时代延续苹果的增长和生态优势。

话题 5:JD Vance Details Trucking Fraud Crackdown Shutting Down 2,000 CDL Mills 链接到标题

  • 分类:AI · Other
  • 概况:热度时间:1 day ago,相关帖子数:46000
  • 是什么事:JD Vance公开谈及整治卡车驾驶证（CDL）造假和“CDL工厂”，称相关执法已关闭约2000家违规机构。
  • 为什么重要:这类监管行动会影响物流与运输行业的数据可信度、合规成本和人才供给，也会间接影响自动驾驶货运、车队管理等AI应用落地时所依赖的安全与身份验证体系。
  • 讨论概况:X上的讨论主要集中在两点：一是支持强力打击欺诈、提升道路安全；二是质疑关闭规模是否过大，以及这会不会进一步加剧卡车司机短缺和行业用工压力。

话题 6:NBA Hits Clippers with Historic Penalties in Kawhi Leonard Cap Probe 链接到标题

  • 分类:AI · Sports
  • 概况:热度时间:1 day ago,相关帖子数:265000
  • 是什么事:NBA因快船队涉嫌在招募和签下科怀·伦纳德过程中规避工资帽，对球队处以创纪录处罚。
  • 为什么重要:该事件本身并非人工智能事件，但体现了体育联盟在复杂数据、合同审查和规则执行中对透明度与治理机制的重视，也可能影响体育分析和商业决策类AI应用所依赖的数据环境。
  • 讨论概况:X上的讨论主要集中在处罚是否与违规程度相称、快船管理层应承担多大责任、联盟调查和执法是否一致，以及这一判罚是否会成为今后处理工资帽规避案件的先例。

话题 7:Liverpool Names 25-Man Champions League Squad, Omits Chiesa and Endo 链接到标题

  • 分类:AI · Other
  • 概况:热度时间:3 hours ago,相关帖子数:4600
  • 是什么事:利物浦公布了25人欧冠报名名单，基耶萨和远藤航未被列入。
  • 为什么重要:这是一则足球阵容新闻，与人工智能技术、产业或研究没有直接关联，主要体现了热度话题分类可能存在跨领域噪声。
  • 讨论概况:X上的讨论焦点主要集中在两名球员落选的原因、球队欧冠阵容取舍及后续出场机会；由于缺少代表性推文，暂无法确认具体舆论倾向。

话题 8:Sophie Cunningham Enjoys Fresh Ground Chuck from Family Friend’s Ranch 链接到标题

  • 分类:AI · Sports
  • 概况:热度时间:23 hours ago,相关帖子数:1800
  • 是什么事:Sophie Cunningham分享或谈及享用来自家族朋友牧场的新鲜牛肉末。
  • 为什么重要:从给定信息看，该话题与AI技术或产业没有直接关联，归入AI分类可能反映了平台热榜的标签偏差或语境识别问题。
  • 讨论概况:材料未提供代表推文，无法可靠判断X上的具体讨论焦点；目前可确认的内容主要围绕球员生活与食品来源展开。

话题 9:1968 NYC Subway Photos Spark Debate on Change and Order 链接到标题

  • 分类:AI · Sports
  • 概况:热度时间:2 hours ago,相关帖子数:405
  • 摘要:1968 NYC Subway Photos Spark Debate on Change and Order:

话题 10:Gabriel Martinelli Joins Al-Hilal in Arsenal’s Record €70m Sale 链接到标题

  • 分类:AI · Sports
  • 概况:热度时间:2 days ago,相关帖子数:192000
  • 摘要:Gabriel Martinelli Joins Al-Hilal in Arsenal’s Record €70m Sale: 🚨🔵⚪️ OFFICIAL: Gabriel Martinelli joins Al Hilal from Arsenal in a €70m package deal. 🇧🇷 It becomes Arsenal’s record sale, with a sell-on clause also included in the agreement.

话题 11:Lisa Reveals Hidden Romances and K-pop Struggles in New Documentary 链接到标题

  • 分类:AI · Entertainment
  • 概况:热度时间:20 hours ago,相关帖子数:61000
  • 摘要:Lisa Reveals Hidden Romances and K-pop Struggles in New Documentary:

话题 12:Elon Musk Documentary Sparks Family Defense and Fierce Backlash 链接到标题

  • 分类:AI · Entertainment
  • 概况:热度时间:1 day ago,相关帖子数:49000
  • 是什么事:一部关于埃隆·马斯克的纪录片引发关注后，其家人出面为他辩护，同时也在 X 上招致大量批评和反弹。
  • 为什么重要:马斯克同时是 AI 公司 xAI 的核心人物，这类舆论事件会影响外界对其个人形象、商业决策和 AI 版图的判断，也会放大围绕 AI 领导者责任与影响力的讨论。
  • 讨论概况:X 上主要在争论纪录片是否公正呈现马斯克、家人辩护是否有说服力，以及马斯克的个人争议是否会连带影响 xAI、特斯拉和相关 AI 议题的公众信任。

话题 13:Maye Musk Rejects Claim of Elon’s Misery from Anonymous Source 链接到标题

  • 分类:AI · Entertainment
  • 概况:热度时间:6 hours ago,相关帖子数:7800
  • 是什么事:Maye Musk公开否认了匿名来源关于“埃隆·马斯克很痛苦/不快乐”的说法，回应了围绕其个人状态的传闻。
  • 为什么重要:马斯克是 xAI、Tesla 和 X 的核心人物，他的个人形象、情绪状态和公众叙事会直接影响外界对其 AI 业务、管理风格和战略判断的关注。
  • 讨论概况:X 上的讨论主要集中在消息来源是否可靠、Maye Musk 的否认是否足以反驳传闻，以及马斯克的个人生活是否会被过度放大并影响对其 AI 事业的评价。

话题 14:Elon Musk Warns Austin Heat Hinders Recruiting 链接到标题

  • 分类:AI · News
  • 概况:热度时间:3 hours ago,相关帖子数:1700
  • 摘要:Elon Musk Warns Austin Heat Hinders Recruiting:

话题 15:Healthcare Worker Leaves Voicemail for Joy as Hope 链接到标题

  • 分类:AI · Entertainment
  • 概况:热度时间:3 hours ago,相关帖子数:7000
  • 是什么事:一名医护人员给名为 Joy 的对象留言，表达希望，这段内容在 X 上引发关注。
  • 为什么重要:该话题体现了 AI 与情感化娱乐内容结合的传播潜力，也引发对合成语音、真实身份和内容真实性的关注。
  • 讨论概况:讨论焦点集中在留言是否由 AI 生成或经过声音合成、故事是否真实，以及这种具有情绪感染力的内容是在传递希望还是利用情感博取流量。

💡 大佬观点(Influencer Insights) 链接到标题

今日大佬观点暂缺,推荐阅读 Watch List 深度内容。

📚 附录:今日 Watch List 更新源列表 链接到标题

时间窗口:最近 3 天;覆盖 22 个源;共 36 条更新

OpenAI Blog (A_full) 链接到标题

  • Daybreak for Frontline Defenders: $1B to protect essential services

    • 发布时间:2026-09-03 21:15 北京时间
    • 摘要:【待翻译】- Today OpenAI is introducing Daybreak for Frontline Defenders, a new global initiative to help frontline defenders use frontier AI cyber capabilities to protect essential services in the United States and around the world.
        • A $1 billion global commitment to expand subsidized access to Daybreak cyber models and products, training, technical support, and partnerships in the United States and internationally.
        • Daybreak for America, bringing together all of OpenAI’s U.S.
      • work to protect the systems Americans rely on every day—from water and electricity to local government and banking—including a new pilot with the Multi-State Information Sharing and Analysis Center (MS-ISAC).
        • More than 35 enterprise products and partner-operated services through the Daybreak Defense Network, bringing Daybreak cyber models into the tools, services, and workflows enterprise defenders already use.
    • EN 要点:
      • OpenAI introduces Daybreak for Frontline Defenders
      • A $1 billion commitment expands access to frontier cyber AI, training, and support for essential services.

  • Legora reviewed 41 documents in minutes with GPT-6 Astra

    • 发布时间:2026-09-03 20:00 北京时间
    • 摘要:【待翻译】- Legora is an agentic operating system for legal and professional work, used by more than 100,000 professionals across more than 1,800 in-house legal departments and law firms in over 50 markets.
      • Its legal engineers work directly with customers to understand how they operate and adapt Legora to their end-to-end workflows, from contract and agreement review to legal research.
      • One of the more tedious workflows is financial-statement tie-out: checking every figure in draft accounts against trial balances, a consolidation schedule, and the previous year’s accounts until each item agrees.
      • As Legora Legal Engineer Percevale Perks says, the work “can take an entire evening, sometimes days.”.

      • Processing complex financial context at scale. 链接到标题

    • EN 要点:
      • Legora used GPT-6 Astra to review 41 documents in minutes, find all four planted errors, and improve performance by nearly 40% in this financial-review workflow…

  • Playco cut manual fixes 50% prototyping games with GPT-6 Astra

    • 发布时间:2026-09-03 20:00 北京时间
    • 摘要:【待翻译】- Using GPT-6 Astra, Playco built three themed game prototypes from one grey box foundation and reported 50% fewer manual fixes than with the previous model.
      • This piece from OpenAI Blog explains how Playco cut manual fixes 50% prototyping games with GPT-6 Astra shapes the broader AI and infrastructure landscape.
      • It also surfaces practical implications for founders, operators, and investors following Playco cut manual fixes 50% prototyping games with GPT-6 Astra.
    • EN 要点:
      • Using GPT-6 Astra, Playco built three themed game prototypes from one grey box foundation and reported 50% fewer manual fixes than with the previous model.

  • Safety overview: GPT-6 Astra

    • 发布时间:2026-09-03 08:00 北京时间
    • 摘要:【待翻译】- GPT-6 Astra is our most capable broadly deployed model and our first to reach the Critical level of cybersecurity capability under our Preparedness Framework.
      • This piece from OpenAI Blog explains how Safety overview: GPT-6 Astra shapes the broader AI and infrastructure landscape.
      • It also surfaces practical implications for founders, operators, and investors following Safety overview: GPT-6 Astra.
    • EN 要点:
      • GPT-6 Astra is our most capable broadly deployed model and our first to reach the Critical level of cybersecurity capability under our Preparedness Framework.

Google DeepMind Blog (A_full) 链接到标题

  • Introducing WeatherNext 3, our most advanced and accurate global weather AI model
    • 发布时间:2026-09-03 23:02 北京时间
    • 摘要:【待翻译】- Introducing WeatherNext 3, our most advanced and accurate global weather AI model.
      • This piece from Google DeepMind Blog explains how Introducing WeatherNext 3, our most advanced and accurate global weather AI model shapes the broader AI and infrastructure landscape.
      • It also surfaces practical implications for founders, operators, and investors following Introducing WeatherNext 3, our most advanced and accurate global weather AI model.
    • EN 要点:
      • Introducing WeatherNext 3, our most advanced and accurate global weather AI model

Two Minute Papers (B_intro+search) 链接到标题

  • Claude Fable AI Is Much Stranger Than The Headlines Suggest
    • 发布时间:2026-09-03 16:23 北京时间
    • 摘要:【待翻译】- ❤️ Check out Lambda here and sign up for their GPU Cloud:.
      • 📝 The Claude Fable 5.1 paper is available here:.
      • Adam Bridges, B Shang, Carlos Galarza, Christian Ahlin, Eric Tyson, Juan Benet, Lukas Biewald, Michael Tedder, Owen Skarpness, Ryan Stankye, Shawn Becker, Steef, Taras Bobrovytsky, Tazaur Sagenclaw, Tybie Fitzhugh, Ueli Gallizzi.
      • Claude Fable AI Is Much Stranger Than The Headlines Suggest.
    • EN 要点:
      • ❤️ Check out Lambda here and sign up for their GPU Cloud:
      • 📝 The Claude Fable 5.1 paper is available here:
      • 🙏 We would like to thank our generous Patreon supporters who make Two Minute Papers possible:
      • Adam Bridges, B Shang, Carlos Galarza, Christian Ahlin, Eric Tyson, Juan Benet, Lukas Biewald, Michael Tedder, Owen Skarpness, Ryan Stankye, Shawn Becker, Steef…

ArXiv cs.AI (B_intro+search) 链接到标题

  • EvalDetectBench: A Benchmark for Measuring Evaluation Awareness in Frontier Language Models

    • 发布时间:2026-09-03 12:00 北京时间
    • 摘要:【待翻译】- arXiv:2609.01611v1 Announce Type: new.
      • Abstract: Frontier large language models can often recognize when they are being evaluated, a capability known as evaluation awareness.
      • If models behave differently in evaluations than in deployment, this undermines the validity of evaluation results, which are a crucial component of current AI safety frameworks.
      • We introduce EvalDetectBench, an open pipeline and benchmark for measuring evaluation awareness that works with any Inspect-compatible evaluation, allowing practitioners to test against current and future benchmarks.
    • EN 要点:
      • arXiv:2609.01611v1 Announce Type: new
      • Abstract: Frontier large language models can often recognize when they are being evaluated, a capability known as evaluation awareness
      • If models behave differently in evaluations than in deployment, this undermines the validity of evaluation results, which are a crucial component of current AI…
      • We introduce EvalDetectBench, an open pipeline and benchmark for measuring evaluation awareness that works with any Inspect-compatible evaluation, allowing prac…

  • Meta-ethics and AI: exploring the novel meta-ethical questions in the era of AI

    • 发布时间:2026-09-03 12:00 北京时间
    • 摘要:【待翻译】- arXiv:2609.01685v1 Announce Type: new.
      • Abstract: With the development of artificial intelligence (AI), the landscape of meta-ethics, which has largely centred on human ethics, faces pressures that may significantly reconfigure it.
      • In particular, if future AI systems were to exhibit sufficiently integrated capacities for moral reasoning, moral intentionality, and moral reflection, novel meta-ethical questions would arise concerning what I call “AI’s own ethics”, as distinct from ethical principles merely imposed on AI by human designers.
      • This paper offers a conditional and methodological framework for identifying the questions that would emerge if such AI systems were to arise.
    • EN 要点:
      • arXiv:2609.01685v1 Announce Type: new
      • Abstract: With the development of artificial intelligence (AI), the landscape of meta-ethics, which has largely centred on human ethics, faces pressures that ma…
      • In particular, if future AI systems were to exhibit sufficiently integrated capacities for moral reasoning, moral intentionality, and moral reflection, novel me…
      • This paper offers a conditional and methodological framework for identifying the questions that would emerge if such AI systems were to arise

  • When Can a Machine Trust a Statute? A Survival Certificate for Machine-Extracted Legal Logic

    • 发布时间:2026-09-03 12:00 北京时间
    • 摘要:【待翻译】- arXiv:2609.01741v1 Announce Type: new.
      • Abstract: Statutes are increasingly parsed by machines before people read them, and the parsers disagree: on Missouri’s statutes, two independently written extractors diverge on numeric-threshold presence at a false-negative rate of 0.43.
      • We ask what formal logic survives such noise.
      • We build a passive survival certificate for the Duquenne-Guigues implication basis of machine-extracted statutory contexts: per-attribute inter-extractor disagreement is measured, replayed against the basis in 1,000 Monte Carlo trials, and an implication is certified only when a one-sided Wilson 95% lower bound on survival reaches 0.95; every certified implication carries premise spans and a minimal counterexample.
    • EN 要点:
      • arXiv:2609.01741v1 Announce Type: new
      • Abstract: Statutes are increasingly parsed by machines before people read them, and the parsers disagree: on Missouri’s statutes, two independently written extr…
      • We ask what formal logic survives such noise
      • We build a passive survival certificate for the Duquenne-Guigues implication basis of machine-extracted statutory contexts: per-attribute inter-extractor disagr…

  • When Does Information Sharing Improve Decentralized Discovery? Aggregation, Independent Rescue, and Equilibrium Selection

    • 发布时间:2026-09-03 12:00 北京时间
    • 摘要:【待翻译】- arXiv:2609.01814v1 Announce Type: new.
      • Abstract: Information sharing can improve a pooled estimate while eliminating independent rescue actions.
      • This paper separates those effects in exact finite discovery models.
      • A centralized action-budget profile shows that equal one-person accuracy can coexist with different portfolio values.
    • EN 要点:
      • arXiv:2609.01814v1 Announce Type: new
      • Abstract: Information sharing can improve a pooled estimate while eliminating independent rescue actions
      • This paper separates those effects in exact finite discovery models
      • A centralized action-budget profile shows that equal one-person accuracy can coexist with different portfolio values

  • Induction and Inquiry via Probabilistic Reasoning over Language and Code

    • 发布时间:2026-09-03 12:00 北京时间
    • 摘要:【待翻译】- arXiv:2609.01815v1 Announce Type: new.
      • Abstract: How humans grow and maintain abstract knowledge from the sparse, streaming noisy data of experience is a longstanding challenge in cognitive science.
      • Any computational account must satisfy at least three desiderata: It must be (1) data-efficient and compute-efficient, (2) capture gradations of uncertainty to support intelligent inquiry and information gathering, and (3) be flexible enough to mentally represent the endless range of concepts people can learn and think about.
      • Here we introduce a computational model that captures these three properties, by encoding symbolic knowledge as mental programs that combine natural language with source code, and sequentially inferring mental programs using LLM-guided Bayesian learning algorithms.
    • EN 要点:
      • arXiv:2609.01815v1 Announce Type: new
      • Abstract: How humans grow and maintain abstract knowledge from the sparse, streaming noisy data of experience is a longstanding challenge in cognitive science
      • Any computational account must satisfy at least three desiderata: It must be (1) data-efficient and compute-efficient, (2) capture gradations of uncertainty to…
      • Here we introduce a computational model that captures these three properties, by encoding symbolic knowledge as mental programs that combine natural language wi…

  • Architecting Conversational Data Systems for Stateless LLM APIs: The Hydration Proxy Pattern

    • 发布时间:2026-09-03 12:00 北京时间
    • 摘要:【待翻译】- arXiv:2609.01834v1 Announce Type: new.
      • Abstract: As enterprise platforms transition to conversational reasoning interfaces, the stateless nature of LLM APIs creates an architectural gap.
      • While statelessness enables horizontal scalability for AI providers, it forces client applications to manage the entire burden of conversational state and semantic memory.
      • The work identifies the Hydration Proxy Pattern, an architecture that decouples session persistence from the reasoning engine.
    • EN 要点:
      • arXiv:2609.01834v1 Announce Type: new
      • Abstract: As enterprise platforms transition to conversational reasoning interfaces, the stateless nature of LLM APIs creates an architectural gap
      • While statelessness enables horizontal scalability for AI providers, it forces client applications to manage the entire burden of conversational state and seman…
      • The work identifies the Hydration Proxy Pattern, an architecture that decouples session persistence from the reasoning engine

  • SSAKG 2.0: An Open-Source Package for Structural Associative Sequence Memory and Context-Based Retrieval

    • 发布时间:2026-09-03 12:00 北京时间
    • 摘要:【待翻译】- arXiv:2609.01849v1 Announce Type: new.
      • Abstract: This article presents SSAKG 2.0, an open-source software package for constructing and operating Structural Sequential Associative Knowledge Graphs (SSAKGs).
      • An SSAKG represents objects as graph vertices and ordered sequences as structural patterns of graph connections.
      • The resulting sparse graph is used as an associative memory in which complete sequences can be reconstructed from a partial, unordered context.
    • EN 要点:
      • arXiv:2609.01849v1 Announce Type: new
      • Abstract: This article presents SSAKG 2.0, an open-source software package for constructing and operating Structural Sequential Associative Knowledge Graphs (SS…
      • An SSAKG represents objects as graph vertices and ordered sequences as structural patterns of graph connections
      • The resulting sparse graph is used as an associative memory in which complete sequences can be reconstructed from a partial, unordered context

  • The Memory Trust Gap: Capability-Dependent Failures in Persistent-Memory Agents

    • 发布时间:2026-09-03 12:00 北京时间
    • 摘要:【待翻译】- arXiv:2609.01852v1 Announce Type: new.
      • Abstract: Persistent memory supports personalized agents, but a stale stored fact can override current authoritative evidence without warning.
      • We study when this harm begins as model capability changes.
      • We evaluate a frozen, closed-set, action-scored benchmark with 2 suites that represent 2 different meanings of “no memory” (a Benefit suite, unsolvable without the stored fact, and a Safety suite, in which an authoritative tool always holds the correct value), on a same-family model-size series (Qwen3 0.6/1.7/4/8B).
    • EN 要点:
      • arXiv:2609.01852v1 Announce Type: new
      • Abstract: Persistent memory supports personalized agents, but a stale stored fact can override current authoritative evidence without warning
      • We study when this harm begins as model capability changes
      • We evaluate a frozen, closed-set, action-scored benchmark with 2 suites that represent 2 different meanings of “no memory” (a Benefit suite, unsolvable without…

  • Belief-Calibrated Optimization: An Explicit World Model for Agentic Optimization

    • 发布时间:2026-09-03 12:00 北京时间
    • 摘要:【待翻译】- arXiv:2609.01861v1 Announce Type: new.
      • Abstract: The performance of an LLM agent depends on the scaffold around a frozen model.
      • A common way to improve that scaffold is to use a coding agent as an optimizer: it reads current scores and traces and iteratively edits the source, producing a new candidate each round.
      • Each edit is chosen according to a belief about how the environment will respond: what went wrong, and which change should help.
    • EN 要点:
      • arXiv:2609.01861v1 Announce Type: new
      • Abstract: The performance of an LLM agent depends on the scaffold around a frozen model
      • A common way to improve that scaffold is to use a coding agent as an optimizer: it reads current scores and traces and iteratively edits the source, producing a…
      • Each edit is chosen according to a belief about how the environment will respond: what went wrong, and which change should help

  • Epistemic Sybil Resistance: Multiplying AI Agents Without Multiplying Evidence

    • 发布时间:2026-09-03 12:00 北京时间
    • 摘要:【待翻译】- arXiv:2609.01873v1 Announce Type: new.
      • Abstract: Multi-agent AI systems improve inference by spawning agents and synthesizing reports.
      • But another agent is not another observation: apparently independent reports may descend from the same evidence, and genuinely independent evidence can produce nearly identical reports.
      • We formalize this as an epistemic Sybil problem.
    • EN 要点:
      • arXiv:2609.01873v1 Announce Type: new
      • Abstract: Multi-agent AI systems improve inference by spawning agents and synthesizing reports
      • But another agent is not another observation: apparently independent reports may descend from the same evidence, and genuinely independent evidence can produce…
      • We formalize this as an epistemic Sybil problem

ArXiv cs.CL (B_intro+search) 链接到标题

  • PRO-Step: Step-level Process Reward Optimization for Retrieval-Augmented Generation

    • 发布时间:2026-09-03 12:00 北京时间
    • 摘要:【待翻译】- arXiv:2609.01658v1 Announce Type: new.
      • Abstract: Retrieval-Augmented Generation enhances Large Language Models by grounding responses in external knowledge, but multi-hop reasoning remains vulnerable to error propagation, where early retrieval failures confound subsequent steps.
      • Standard outcome-based optimization only rewards the final answer, leaving intermediate retrieval and reasoning errors undetected.
      • While existing process-based methods introduce step-level signals, they still score each step against the final answer, rewarding spurious successes where flawed retrieval coincidentally produces the correct answer.
    • EN 要点:
      • arXiv:2609.01658v1 Announce Type: new
      • Abstract: Retrieval-Augmented Generation enhances Large Language Models by grounding responses in external knowledge, but multi-hop reasoning remains vulnerable…
      • Standard outcome-based optimization only rewards the final answer, leaving intermediate retrieval and reasoning errors undetected
      • While existing process-based methods introduce step-level signals, they still score each step against the final answer, rewarding spurious successes where flawe…

  • Learning Evidence Sufficiency Boundaries for Selective Answering in Grounded Multi-Hop QA

    • 发布时间:2026-09-03 12:00 北京时间
    • 摘要:【待翻译】- arXiv:2609.01687v1 Announce Type: new.
      • Abstract: Grounded question answering systems should answer only when the supplied evidence supports the answer.
      • In multi-hop QA, this requirement is difficult because partial evidence can make an unsupported answer appear plausible.
      • We study selective answering through evidence sufficiency boundaries: for the same question, a model should abstain under unsupported or partially supported context, answer when the context first becomes sufficient, and keep the answer stable when redundant evidence is added.
    • EN 要点:
      • arXiv:2609.01687v1 Announce Type: new
      • Abstract: Grounded question answering systems should answer only when the supplied evidence supports the answer
      • In multi-hop QA, this requirement is difficult because partial evidence can make an unsupported answer appear plausible
      • We study selective answering through evidence sufficiency boundaries: for the same question, a model should abstain under unsupported or partially supported con…

  • SpeakPay: Domain-Adaptive LoRA Fine-Tuning of Whisper for Low-Resource Nepali Financial Speech Recognition

    • 发布时间:2026-09-03 12:00 北京时间
    • 摘要:【待翻译】- arXiv:2609.01737v1 Announce Type: new.
      • Abstract: Mobile payment applications in Nepal are graphically mediated and largely inaccessible to visually impaired users.
      • This paper presents SpeakPay, a voice-first digital wallet, and documents the central technical contribution: a controlled study of domain adaptation for low-resource financial speech recognition.
      • We introduce NepFinSpeech-403, a 403-utterance dataset of Nepali financial voice commands (send, load, and balance operations spanning 237 unique numerals), and fine-tune Whisper large-v2 with LoRA.
    • EN 要点:
      • arXiv:2609.01737v1 Announce Type: new
      • Abstract: Mobile payment applications in Nepal are graphically mediated and largely inaccessible to visually impaired users
      • This paper presents SpeakPay, a voice-first digital wallet, and documents the central technical contribution: a controlled study of domain adaptation for low-re…
      • We introduce NepFinSpeech-403, a 403-utterance dataset of Nepali financial voice commands (send, load, and balance operations spanning 237 unique numerals), and…

  • MemeCULT-1K: Benchmarking South Asian Cultural Context and Humor Understanding of Multimodal Models

    • 发布时间:2026-09-03 12:00 北京时间
    • 摘要:【待翻译】- arXiv:2609.01772v1 Announce Type: new.
      • Abstract: Meme understanding goes beyond recognizing visual content or literal text; it requires implicit cultural knowledge and pragmatic inference that most vision-language models still lack.
      • We introduce MemeCULT-1K, a multilingual benchmark of 1,000 South Asian memes in Bengali, English, and Hindi, where each meme is paired with a cultural context note and three human-written explanations, along with a supplementary set of 54 Bengali regional dialect memes.
      • We evaluate thirteen popular Vision Language Models (VLMs) under two settings: meme-only and context-aware.
    • EN 要点:
      • arXiv:2609.01772v1 Announce Type: new
      • Abstract: Meme understanding goes beyond recognizing visual content or literal text; it requires implicit cultural knowledge and pragmatic inference that most v…
      • We introduce MemeCULT-1K, a multilingual benchmark of 1,000 South Asian memes in Bengali, English, and Hindi, where each meme is paired with a cultural context…
      • We evaluate thirteen popular Vision Language Models (VLMs) under two settings: meme-only and context-aware

  • VakyArth: Evaluating Pragmatic Competence in LLMs across Indic Languages

    • 发布时间:2026-09-03 12:00 北京时间
    • 摘要:【待翻译】- arXiv:2609.01788v1 Announce Type: new.
      • Abstract: Real-world communication often requires pragmatic reasoning: interpreting meanings implied through context and cultural convention rather than stated literally.
      • Existing pragmatic evaluation remains largely limited to English and high-resource languages, leaving Indic languages unexplored despite their linguistic and cultural diversity.
      • We introduce VakyArth, the first pragmatic benchmark for Indic languages, designed as a diagnostic evaluation covering Hindi, Punjabi, Tamil, and Malayalam.
    • EN 要点:
      • arXiv:2609.01788v1 Announce Type: new
      • Abstract: Real-world communication often requires pragmatic reasoning: interpreting meanings implied through context and cultural convention rather than stated…
      • Existing pragmatic evaluation remains largely limited to English and high-resource languages, leaving Indic languages unexplored despite their linguistic and cu…
      • We introduce VakyArth, the first pragmatic benchmark for Indic languages, designed as a diagnostic evaluation covering Hindi, Punjabi, Tamil, and Malayalam

  • Disentangling Statistical Preemption from Entrenchment in Language Models’ Avoidance of Overgeneralization

    • 发布时间:2026-09-03 12:00 北京时间
    • 摘要:【待翻译】- arXiv:2609.01794v1 Announce Type: new.
      • Abstract: How do learners avoid overgeneralizations such as Tom laughed me without explicit negative evidence?
      • Constructionists have posited two proposals that describe indirect negative evidence against overgeneralizations: preemption (which privileges exposure to near-synonymous construction—e.g., she made him laugh) vs.
      • entrenchment (all exposures to a verb’s grammatical usages, including cases like He laughed).
    • EN 要点:
      • arXiv:2609.01794v1 Announce Type: new
      • Abstract: How do learners avoid overgeneralizations such as Tom laughed me without explicit negative evidence
      • Constructionists have posited two proposals that describe indirect negative evidence against overgeneralizations: preemption (which privileges exposure to near-…
      • entrenchment (all exposures to a verb’s grammatical usages, including cases like He laughed)

  • How Do Prompt Variations Affect Energy Consumption in On-Device LLMs?

    • 发布时间:2026-09-03 12:00 北京时间
    • 摘要:【待翻译】- arXiv:2609.01798v1 Announce Type: new.
      • Abstract: Large language models (LLMs) are increasingly deployed on mobile devices, making energy efficiency a key deployment constraint, yet the energy impact of prompt design remains underexplored.
      • This paper aims to understand how two prompt properties, cognitive load and phrasing pattern, shape the energy behavior of on-device LLM inference.
      • We conduct a broad empirical study covering prompt properties, datasets, models, and devices, with phase-level profiling that separates prefill and decode energy.
    • EN 要点:
      • arXiv:2609.01798v1 Announce Type: new
      • Abstract: Large language models (LLMs) are increasingly deployed on mobile devices, making energy efficiency a key deployment constraint, yet the energy impact…
      • This paper aims to understand how two prompt properties, cognitive load and phrasing pattern, shape the energy behavior of on-device LLM inference
      • We conduct a broad empirical study covering prompt properties, datasets, models, and devices, with phase-level profiling that separates prefill and decode energ…

  • TalkFa: A Unified Benchmark for Farsi Dialogue Generation and Understanding

    • 发布时间:2026-09-03 12:00 北京时间
    • 摘要:【待翻译】- arXiv:2609.01810v1 Announce Type: new.
      • Abstract: Farsi, spoken by more than 120 million people, lacks a comprehensive benchmark for dialogue generation and understanding.
      • We introduce TALKFA, a unified benchmark comprising three complementary datasets: (1) WIKI-FADIAL, 4.2K Wikipedia-grounded dialogues for knowledge-grounded generation; (2) DAILYDIALOG-FA, 6.6K dialogues annotated for dialogue acts and emotions; and (3) PLAYDIAL-FA, 2.1K theatrical dialogues with sentiment labels.
      • While LLMs assist data construction, every dialogue undergoes multi-stage review and revision by native Farsi speakers, and only the final human-approved dialogues are released.
    • EN 要点:
      • arXiv:2609.01810v1 Announce Type: new
      • Abstract: Farsi, spoken by more than 120 million people, lacks a comprehensive benchmark for dialogue generation and understanding
      • We introduce TALKFA, a unified benchmark comprising three complementary datasets: (1) WIKI-FADIAL, 4.2K Wikipedia-grounded dialogues for knowledge-grounded gene…
      • While LLMs assist data construction, every dialogue undergoes multi-stage review and revision by native Farsi speakers, and only the final human-approved dialog…

  • AVERT: Audio-Verified Adjudication for Spoken Dialogue State Tracking

    • 发布时间:2026-09-03 12:00 北京时间
    • 摘要:【待翻译】- arXiv:2609.01828v1 Announce Type: new.
      • Abstract: Spoken dialogue state tracking recovers slot-value pairs from speech, where ASR errors concentrate in entity values and persist across turns, making it both a generation and an editing problem.
      • A strong per-turn text editor corrects much of this but, operating on the transcript alone, leaves three recoverable errors: a value predicted inconsistently across turns, an omitted slot, and a value the audio does not support.
      • We present AVERT, which scores each candidate value by combining cross-turn agreement with a trained audio-conditioned verifier and resolves the three error types with three operators, vote, add, and swap, each restricted to the slots where its error is common.
    • EN 要点:
      • arXiv:2609.01828v1 Announce Type: new
      • Abstract: Spoken dialogue state tracking recovers slot-value pairs from speech, where ASR errors concentrate in entity values and persist across turns, making i…
      • A strong per-turn text editor corrects much of this but, operating on the transcript alone, leaves three recoverable errors: a value predicted inconsistently ac…
      • We present AVERT, which scores each candidate value by combining cross-turn agreement with a trained audio-conditioned verifier and resolves the three error typ…

  • Interpretable Symptom Vectors for Depression in a Large Language Model

    • 发布时间:2026-09-03 12:00 北京时间
    • 摘要:【待翻译】- arXiv:2609.01832v1 Announce Type: new.
      • Abstract: Patients with depression present with diverse symptom profiles, yet clinical practice routinely reduces this variation to a single severity score.
      • Large language models (LLMs) can potentially capture various symptoms and their severity from patient speech.
      • However, how depressive symptoms are represented inside LLMs remains poorly understood, limiting clinical trust.
    • EN 要点:
      • arXiv:2609.01832v1 Announce Type: new
      • Abstract: Patients with depression present with diverse symptom profiles, yet clinical practice routinely reduces this variation to a single severity score
      • Large language models (LLMs) can potentially capture various symptoms and their severity from patient speech
      • However, how depressive symptoms are represented inside LLMs remains poorly understood, limiting clinical trust

ArXiv cs.LG (B_intro+search) 链接到标题

  • WMLLM: Self-Evolving Optimization Agents via Predict-Then-Act World Modeling

    • 发布时间:2026-09-03 12:00 北京时间
    • 摘要:【待翻译】- arXiv:2609.01608v1 Announce Type: new.
      • Abstract: Black-box optimization problems remain challenging because of large, weakly structured, and high-dimensional search spaces.
      • Existing methods often suffer from poor sample efficiency because they rely on direct candidate generation or trial-and-error refinement.
      • A natural way to improve search efficiency is to use world modeling, which can help identify promising optimization directions before costly evaluation.
    • EN 要点:
      • arXiv:2609.01608v1 Announce Type: new
      • Abstract: Black-box optimization problems remain challenging because of large, weakly structured, and high-dimensional search spaces
      • Existing methods often suffer from poor sample efficiency because they rely on direct candidate generation or trial-and-error refinement
      • A natural way to improve search efficiency is to use world modeling, which can help identify promising optimization directions before costly evaluation

  • DiDrive: A Risk-Aware Hierarchical Diffusion Framework for Safe Offline Reinforcement Learning in Autonomous Driving

    • 发布时间:2026-09-03 12:00 北京时间
    • 摘要:【待翻译】- arXiv:2609.01609v1 Announce Type: new.
      • Abstract: While diffusion models effectively capture multimodal behavioral priors for autonomous driving, offline reinforcement learning (RL) policies remain susceptible to distribution shift, heavy-tailed risk signals, out-of-distribution (OOD) action generation, and high-dimensional state redundancy.
      • To address these challenges, we propose DiDrive, a distribution-guided offline diffusion framework featuring two synergistic components: the Risk-Aware Hierarchical Diffusion (RHDif) architecture and the 3DICE policy optimization paradigm.
      • In the state space, RHDif utilizes a low-level risk-gated encoder and a high-level contextual modulator to filter environmental redundancy and focus on safety-critical threats.
    • EN 要点:
      • arXiv:2609.01609v1 Announce Type: new
      • Abstract: While diffusion models effectively capture multimodal behavioral priors for autonomous driving, offline reinforcement learning (RL) policies remain su…
      • To address these challenges, we propose DiDrive, a distribution-guided offline diffusion framework featuring two synergistic components: the Risk-Aware Hierarch…
      • In the state space, RHDif utilizes a low-level risk-gated encoder and a high-level contextual modulator to filter environmental redundancy and focus on safety-c…

  • Prompt-Space Meta-Learning Does Not Transfer Across Users: A Frozen-LLM Negative Result

    • 发布时间:2026-09-03 12:00 北京时间
    • 摘要:【待翻译】- arXiv:2609.01615v1 Announce Type: new.
      • Abstract: Personalizing a frozen large language model (LLM) to individual users is often framed as a meta-learning problem in prompt space: each user is a task, and one seeks a shared natural-language adaptation policy that, given a handful of the user’s labeled interactions, configures the frozen model for that user.
      • The framing is attractive because it is backbone-agnostic and reuses the machinery of prompt optimization, yet the field rarely tests whether the optimized meta-objective encodes transferable cross-user adaptation rather than generic instruction quality.
      • We study this question with Muse (Meta-learned User-adaptation via Shared Evolution), which evolves a single shared adaptation prompt over a meta-train user population by reflective prompt evolution, freezes it, and applies it zero-shot to held-out users; matched controls isolate learning from confounds of phrasing and selection.
    • EN 要点:
      • arXiv:2609.01615v1 Announce Type: new
      • Abstract: Personalizing a frozen large language model (LLM) to individual users is often framed as a meta-learning problem in prompt space: each user is a task,…
      • The framing is attractive because it is backbone-agnostic and reuses the machinery of prompt optimization, yet the field rarely tests whether the optimized meta…
      • We study this question with Muse (Meta-learned User-adaptation via Shared Evolution), which evolves a single shared adaptation prompt over a meta-train user pop…

  • Efficient Context-Limited Telescope Bibliography Classification for the WASP-2025 Shared Task Using SciBERT

    • 发布时间:2026-09-03 12:00 北京时间
    • 摘要:【待翻译】- arXiv:2609.01647v1 Announce Type: new.
      • Abstract: The creation of telescope bibliographies is a crucial part of assessing the scientific impact of observatories and ensuring reproducibility in astronomy.
      • This task involves identifying, categorizing, and linking scientific publications that reference or use specific telescopes.
      • However, this process remains largely manual and resource intensive.
    • EN 要点:
      • arXiv:2609.01647v1 Announce Type: new
      • Abstract: The creation of telescope bibliographies is a crucial part of assessing the scientific impact of observatories and ensuring reproducibility in astrono…
      • This task involves identifying, categorizing, and linking scientific publications that reference or use specific telescopes
      • However, this process remains largely manual and resource intensive

  • CliffRank: A Dual-Branch Framework for Activity-Cliff Ranking Prediction

    • 发布时间:2026-09-03 12:00 北京时间
    • 摘要:【待翻译】- arXiv:2609.01673v1 Announce Type: new.
      • Abstract: Activity-cliff ranking remains difficult because local structural changes can cause large activity differences, while high-quality data that resolve the underlying mechanisms remain limited.
      • To use available activity labels more effectively, we combine absolute-activity regression with ranking-consistency learning.
      • CliffRank trains two parallel predictors with mean squared error, a thresholded listwise loss, and Pairwise Preference Consistency (PPC), which aligns relative ordering in the preference-probability space.
    • EN 要点:
      • arXiv:2609.01673v1 Announce Type: new
      • Abstract: Activity-cliff ranking remains difficult because local structural changes can cause large activity differences, while high-quality data that resolve t…
      • To use available activity labels more effectively, we combine absolute-activity regression with ranking-consistency learning
      • CliffRank trains two parallel predictors with mean squared error, a thresholded listwise loss, and Pairwise Preference Consistency (PPC), which aligns relative…

  • Sim2Signal: Sim-to-Real Benchmarks for Traffic Signal Control

    • 发布时间:2026-09-03 12:00 北京时间
    • 摘要:【待翻译】- arXiv:2609.01676v1 Announce Type: new.
      • Abstract: Reinforcement learning achieves strong traffic signal control performance in simulation, yet policies trained in simulators often fail once deployed in the real world, a failure known as the Sim-to-Real gap.
      • When RL is applied to traffic signal control, this gap arises from several sources: sensing, action execution, traffic dynamics, and the control objective.
      • Their relative impact and the reliability of existing Sim-to-Real mitigation methods remain insufficiently understood, and the field lacks a standard benchmark for systematically measuring the gap and evaluating mitigation methods.
    • EN 要点:
      • arXiv:2609.01676v1 Announce Type: new
      • Abstract: Reinforcement learning achieves strong traffic signal control performance in simulation, yet policies trained in simulators often fail once deployed i…
      • When RL is applied to traffic signal control, this gap arises from several sources: sensing, action execution, traffic dynamics, and the control objective
      • Their relative impact and the reliability of existing Sim-to-Real mitigation methods remain insufficiently understood, and the field lacks a standard benchmark…

  • A Survey on Self-Improving Test-Time Intelligence: Feedback-Driven Adapting, Learning, and Scaling at Inference

    • 发布时间:2026-09-03 12:00 北京时间
    • 摘要:【待翻译】- arXiv:2609.01679v1 Announce Type: new.
      • Abstract: The ability of AI systems to improve their behavior during deployment is becoming increasingly important.
      • As inference moves beyond the static execution of a fixed trained model, a growing body of work studies how models can refine their behavior on the fly by exploiting test-time information and additional computation.
      • These developments have largely evolved along two directions: methods that modify the model’s state using test-time signals, and methods that improve predictions through extra inference-time resources such as more sampling and tool use.
    • EN 要点:
      • arXiv:2609.01679v1 Announce Type: new
      • Abstract: The ability of AI systems to improve their behavior during deployment is becoming increasingly important
      • As inference moves beyond the static execution of a fixed trained model, a growing body of work studies how models can refine their behavior on the fly by explo…
      • These developments have largely evolved along two directions: methods that modify the model’s state using test-time signals, and methods that improve prediction…

  • Reinforcement Learning and Rule-Based Peer-to-Peer Pricing in Residential PV-BES Communities

    • 发布时间:2026-09-03 12:00 北京时间
    • 摘要:【待翻译】- arXiv:2609.01680v1 Announce Type: new.
      • Abstract: This paper compares rule-based and learning-based pricing mechanisms for peer-to-peer (P2P) electricity trading in residential photovoltaic communities.
      • The rule-based benchmarks comprise bill-sharing as an ex post allocation mechanism, the mid-market rate, and supply-demand-ratio pricing.
      • The reinforcement-learning (RL) formulation is implemented through a Deep Q-Network and evaluated under multiplier-based and learnable SDR-shaped pricing, with a fixed-parameter SDR variant as a non-learning control.
    • EN 要点:
      • arXiv:2609.01680v1 Announce Type: new
      • Abstract: This paper compares rule-based and learning-based pricing mechanisms for peer-to-peer (P2P) electricity trading in residential photovoltaic communitie…
      • The rule-based benchmarks comprise bill-sharing as an ex post allocation mechanism, the mid-market rate, and supply-demand-ratio pricing
      • The reinforcement-learning (RL) formulation is implemented through a Deep Q-Network and evaluated under multiplier-based and learnable SDR-shaped pricing, with…

  • Median-of-Means as an Extremal Convex Estimator and a Nonconvex Route to the Trimmed Oracle

    • 发布时间:2026-09-03 12:00 北京时间
    • 摘要:【待翻译】- arXiv:2609.01689v1 Announce Type: new.
      • Abstract: We revisit median-of-means estimation from a deterministic optimization viewpoint and develop a family of block-Lp estimators for robust learning with heavy-tailed and adversarially corrupted data.
      • In a block contamination model with at least a fraction 1 minus epsilon of good blocks, we first show that every convex block M-estimator has worst-case robustness constant at least 1 divided by 1 minus 2 epsilon.
      • This matches the classical median-of-means bound and proves that the trimmed-block oracle constant 1 divided by 1 minus epsilon cannot be attained within the convex class.
    • EN 要点:
      • arXiv:2609.01689v1 Announce Type: new
      • Abstract: We revisit median-of-means estimation from a deterministic optimization viewpoint and develop a family of block-Lp estimators for robust learning with…
      • In a block contamination model with at least a fraction 1 minus epsilon of good blocks, we first show that every convex block M-estimator has worst-case robustn…
      • This matches the classical median-of-means bound and proves that the trimmed-block oracle constant 1 divided by 1 minus epsilon cannot be attained within the co…

  • Tri-Band Channel Measurement-Enabled Multi-Layer Digital Twin for Terahertz Wireless Data Centers

    • 发布时间:2026-09-03 12:00 北京时间
    • 摘要:【待翻译】- arXiv:2609.01699v1 Announce Type: new.
      • Abstract: The rapid growth of AI computing has driven increasing demands for flexible and high-capacity data-center interconnections.
      • Owing to its ultra-wide bandwidth and high spatial reuse capability, terahertz (THz) communication has emerged as a promising solution for future wireless data centers, while digital twins (DTs) enable efficient wireless planning and real-time optimization.
      • In this work, a measurement-driven multi-layer DT framework is proposed for THz wireless data centers, where the physical, channel, evaluation, and manipulation layers are progressively constructed from bottom to top.
    • EN 要点:
      • arXiv:2609.01699v1 Announce Type: new
      • Abstract: The rapid growth of AI computing has driven increasing demands for flexible and high-capacity data-center interconnections
      • Owing to its ultra-wide bandwidth and high spatial reuse capability, terahertz (THz) communication has emerged as a promising solution for future wireless data…
      • In this work, a measurement-driven multi-layer DT framework is proposed for THz wireless data centers, where the physical, channel, evaluation, and manipulation…