AAI Search
AI 资讯日报 · 2026-09-15

2026-09-15 · AI 资讯日报

今日新收录 141 条公开资讯，按模型 / 产品 / 行业 / 论文 / 观点 自动归类汇编（非 AI 生成，点击可溯源原文）。

今日精选6 条

  1. 1.
    论文研究Type Diversity Enables Transformers to Generalise Compositionally— arXiv

    Compositional generalisation has been divided into lexical and structural generalisation. Previous work has found that structural generalisation is harder than lexical for Transformers. We propose tha…

  2. 2.
    论文研究Rethinking Heterogeneous System Disaggregation for Subquadratic Attention— arXiv

    Frontier language models are more aggressively using subquadratic attention to reduce the memory footprint and compute requirements during inference while still delivering frontier accuracy. While exi…

  3. 3.
    论文研究A Hybrid LSTM-XGBoost Framework for Multi-Horizon Stock Return Prediction Across Diversified Equity Portfolios— arXiv

    Accurate prediction of equity returns remains a major challenge in computational finance due to the non-stationary, nonlinear, and low signal-to-noise ratio nature of financial time series. This paper…

  4. 4.
    论文研究CMA-OT: Hierarchical Expert Supervision for Dance-to-Music Generation— arXiv

    Dance-to-music (D2M) generation aims to synthesize music that is rhythmically and stylistically aligned with dance videos. A key challenge arises from the semantic mismatch between sparse dance cues,…

  5. 5.
    论文研究Continue, Adapt, or Yield: In-Turn Adaptation to Overlapping Speech in Full-Duplex Agents— arXiv

    Full-duplex evaluation often emphasizes whether an agent keeps speaking or stops. That binary cannot express a third response humans use routinely: continuing to speak while incorporating what the lis…

  6. 6.
    论文研究A Ranking Approach for Measuring Calibration— arXiv

    When providing forecasted probabilities with a predictive model, the ideal model offers perfect calibration: the true probability of the outcome (i.e., the probability that $Y=1$) exactly matches the…

模型发布/更新6 条

  1. 1.
    Inside Claude for Financial Advisors— Anthropic

    AI 导读 · AI开始切入持牌顾问场景，金融合规与模型可信度将成新战场。

  2. 2.
    How Fyxer built an AI executive assistant people trust— OpenAI

    Fyxer uses OpenAI models, fine-tuning, memory, and real user feedback to organize inboxes and draft emails in each user’s voice.

  3. 3.
    Perplexity Portable Computer Is Now Available on Windows, Powered by NVIDIA RTX— NVIDIA

    As local models become more capable, AI agents can handle more work directly on a PC while keeping sensitive information on the device. Portable Computer is a local version of the…

  4. 4.
    Watch astronaut Christina Koch and Google’s James Manyika discuss space, technology, and discovery.— Google AI

    AI 导读 · 谷歌研究高管对话女航天员，科技如何助力深空探索值得关注。

  5. 5.
    DevFest is back— Google AI

    DevFest 2026 is back and here’s how you can connect with one of the more than 800 global events to build, secure, and scale in the agentic AI era.

  6. 6.
    Anthropic Claude Opus 5.2 灰度测试曝光：响应更快，告别“偷懒”— IT之家

    AI 导读 · 响应提速与不再敷衍，暗示大模型正从堆参数转向拼实用体验，编程场景或率先受益。

产品发布/更新6 条

  1. 7.
    Fashion app Daydream uses Apple Intelligence to help you shop the outfits in your camera roll— TechCrunch

    AI 导读 · 苹果智能让相册穿搭照直接变购物清单，时尚电商的入口正从搜索框移向系统级AI。

  2. 8.
    youngyangyang04/llm-master— GitHub

    大模型（LLM）全栈学习路线与中文教程🔥：覆盖 Prompt Engineering、RAG、AI Agent、MCP、微调、模型部署、Transformer、AI 编程与大厂面试，从入门到生产实践。

  3. 9.
    shadcn-ui/lint— GitHub

    An agent-first linter for Tailwind design systems. Write design system rules that agents can verify.

  4. 10.
    bojieli/ai-infra-book— GitHub

    《深入理解 AI Infra：量化分析与系统设计》（李博杰 著）开源书稿：从硬件约束和模型架构出发，量化推导 LLM 推理与训练系统设计。含全书正文、PDF、配套计算工具与实验

  5. 11.
    sam70361/aora-bot— GitHub

    Emotion Ball 是一套面向 AI 助手的表情引擎:32 种状态表情全部由纯 SVG 与原生 JavaScript 实时驱动,零框架、零图片资源。AI 侧只需输出一个 emotionId,小球即可切换到对应表情,可直接用作聊天机器人、桌面宠物、悬浮助手的情绪表达层。

  6. 12.
    CG3M 版微星 Claw 8 EX AI+ 掌机曝光：10 核 Xe3，最高 90 INT8 TOPS— IT之家

    AI 导读 · 英特尔新一代掌机芯片首秀，Xe3 架构加高算力，掌上游戏与本地 AI 的融合值得盯。

行业动态6 条

  1. 13.
    GPT-5.6 Luna vs. GPT-6 Astra: Is a $1.20 Model Good Enough for Code Review?— Hacker News

    AI 导读 · 价格战打向代码审查，实用性与成本的平衡才是企业真正看点。

  2. 14.
    OpenAI bots knew about the RubyGems caching vulnerability— Hacker News
  3. 15.
    Big AI sets out its terms for regulatory capture— Hacker News
  4. 16.
    For AI leaders Doom is a form of hype— Hacker News
  5. 17.
    Only at TechCrunch Disrupt 2026: What happens when OpenAI ships your roadmap?— TechCrunch

    If you're building an AI company, the question isn't whether foundation models will continue to evolve. It's whether your company will continue creating value as they do. Don't mis…

  6. 18.
    Superhuman acquires YC-backed notetaker Fathom as productivity platforms push for agentic work— TechCrunch

    The notetaker offers a generous free plan, and that has resulted in over 400,000 monthly active users. The company said that over 1 million people have recorded meetings until now.

论文研究6 条

  1. 19.
    Type Diversity Enables Transformers to Generalise Compositionally— arXiv

    Compositional generalisation has been divided into lexical and structural generalisation. Previous work has found that structural generalisation is harder than lexical for Transformers. We propose tha…

  2. 20.
    Rethinking Heterogeneous System Disaggregation for Subquadratic Attention— arXiv

    Frontier language models are more aggressively using subquadratic attention to reduce the memory footprint and compute requirements during inference while still delivering frontier accuracy. While exi…

  3. 21.
    A Hybrid LSTM-XGBoost Framework for Multi-Horizon Stock Return Prediction Across Diversified Equity Portfolios— arXiv

    Accurate prediction of equity returns remains a major challenge in computational finance due to the non-stationary, nonlinear, and low signal-to-noise ratio nature of financial time series. This paper…

  4. 22.
    CMA-OT: Hierarchical Expert Supervision for Dance-to-Music Generation— arXiv

    Dance-to-music (D2M) generation aims to synthesize music that is rhythmically and stylistically aligned with dance videos. A key challenge arises from the semantic mismatch between sparse dance cues,…

  5. 23.
    Continue, Adapt, or Yield: In-Turn Adaptation to Overlapping Speech in Full-Duplex Agents— arXiv

    Full-duplex evaluation often emphasizes whether an agent keeps speaking or stops. That binary cannot express a third response humans use routinely: continuing to speak while incorporating what the lis…

  6. 24.
    A Ranking Approach for Measuring Calibration— arXiv

    When providing forecasted probabilities with a predictive model, the ideal model offers perfect calibration: the true probability of the outcome (i.e., the probability that $Y=1$) exactly matches the…

技巧与观点6 条

  1. 25.
    Abnormal AI: Amazon Bedrock AgentCore for agentic email security at scale— AWS ML

    AI 导读 · AI智能体借临时计算沙箱处理海量邮件威胁，为安全自动化落地提供可复用范式。

  2. 26.
    Manage end-user OAuth consent for AI agents with Amazon Bedrock AgentCore— AWS ML

    AI 导读 · AI代理代用户授权长期缺乏可控方案，亚马逊把OAuth同意流程托管化，有望成为智能体身份治理的行业标

  3. 27.
    How Ninth Wave built AI-powered open finance onboarding on Amazon Bedrock— AWS ML

    Learn how Ninth Wave built Compass, a multi-agent AI onboarding assistant on Amazon Bedrock AgentCore that validates bank APIs against Financial Data Exchange (FDX) standards, scor…

  4. 28.
    The generative AI customization spectrum: From prompt engineering to custom models on AWS— AWS ML

    Pick the right generative AI customization approach on AWS with an 8-step decision framework, from prompt engineering and RAG to fine-tuning, continued pre-training, and Amazon Nov…

  5. 29.
    Automate replenishment with MMF, Databricks Genie, and Amazon Quick— AWS ML

    Foundation models made catalog-wide demand forecasting easy; the hard part is now acting on the forecast. This post builds a closed detect-decide-act loop on Databricks and Amazon…

  6. 30.
    Async GRPO with LoRA across HF Jobs: a bucket, a proxy, and no NCCL— HuggingFace Blog

快讯

日报生成时间：2026-09-28
← 2026-09-14最新日报2026-09-16 →