AAI Search
AI 资讯日报 · 2026-07-28

2026-07-28 · AI 资讯日报

今日新收录 59 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Certified Parallel-in-Time Sinkhorn for Dynamic Entropic Optimal Transport— arXiv

    Dynamic applications, including optimal-transport Flow Matching, repeatedly solve related entropic optimal transport problems, yet conventional distributed Sinkhorn processes frames sequentially and s…

  2. 2.
    论文研究Learning Distributions from Multiple Data Providers— arXiv

    Motivated by learning from heterogeneous and overlapping data providers, we study a stylized model of distribution learning from restricted conditional samples. The goal is to learn an unknown distrib…

  3. 3.
    论文研究KANEx: Translating Kolmogorov-Arnold Networks' Interpretability to Medical Explainability— arXiv

    Computer vision models have become highly effective for medical applications, yet their black-box nature continues to undermine clinician trust. In clinical workflows, chest X-ray classifiers are incr…

  4. 4.
    论文研究Global Convergence of DGM and PINN Algorithms for Solving Nonlinear PDEs— arXiv

    The Deep Galerkin Method (DGM) and Physics Informed Neural Networks (PINNs) have become widely-used methods for solving partial differential equations (PDEs) in the rapidly growing field of scientific…

  5. 5.
    论文研究The Physics of Multi-Turn Long-Horizon Planning: From Pre-training to Post-training via Single- and Multi-Teacher On-Policy Agentic Distillation— arXiv

    Multi-turn long-horizon planning is critical for foundation model agents, yet how to fundamentally improve it remains unclear. Existing models are trained on uncontrollable and opaque Internet data, m…

  6. 6.
    论文研究DataOrchestra: Learning to Orchestrate Per-Example Curation of Pretraining Data— arXiv

    Pretraining data processing is critical to the downstream performance of Large Language Models (LLMs). However, many existing approaches define a fixed processing strategy at the corpus or domain leve…

模型发布/更新5 条

  1. 1.
    Reimagining Independence: How Meta’s AI Models Are Helping the University of Pittsburgh Transform Assistive Robotics— Meta AI

    Reimagining Independence: How Meta’s AI Models Are Helping the University of Pittsburgh Transform Assistive Robotics AI at Meta

  2. 2.
    moonshotai/Kimi-K3— Simon Willison

    AI 导读 · 开源2.8万亿参数模型，体量巨大，预示国产AI大模型新高度。

  3. 3.
    How Claude Performs on Robotics Tasks— Anthropic

    How Claude Performs on Robotics Tasks Anthropic

  4. 4.
    灵光App上线“一键部署”— 36氪

    AI 导读 · 降低AI应用部署门槛，打通创作到分发的全链路，推动全民开发时代加速到来。

  5. 5.
    派早报：月之暗面开源 Kimi K3 模型等— 少数派

    AI 导读 · 月之暗面开源K3模型，可能推动国产大模型技术生态开放与创新。

产品发布/更新6 条

  1. 6.
    sunyifeng11111/sun-felt-wechat-illustrator— GitHub

    为中文公众号文章生成真实 3D 毛毡质感的封面、正文配图与精确插入指南的 Codex Skill。

  2. 7.
    智能体走向终端，个人AI时代正在到来— 量子位

    AI 导读 · 终端落地加速，个人AI时代门槛降低，应用场景即将爆发。

  3. 8.
    baldaworks/callee— GitHub

    Markdown-defined provider-backed agents and deterministic workflows for ACP runtimes.

  4. 9.
    S40911120/recensa— GitHub

    Self-hosted web viewer for Claude Code session transcripts — read, search, replay, and audit every session you have ever run

  5. 10.
    cowboycodr/sitegeist— GitHub

    A visual benchmark testing whether leading coding agents repeat the same design patterns across 100 neutral website briefs.

  6. 11.
    HunterLzap/rag-tender— GitHub

    面向招投标场景的本地化 RAG 标书解析、资质匹配与红线提醒助手

行业动态6 条

  1. 12.
    Anthropic’s Dario Amodei responds: doesn’t oppose open-weight models, but fears Chinese AI— TechCrunch

    AI 导读 · Anthropic CEO表态担忧中国AI，折射中美科技竞争新焦点。

  2. 13.
    英伟达重构工程软件栈：Agent、PhysicsNeMo 与 CUDA-X 合流— InfoQ

    AI 导读 · 英伟达统一三大技术栈，加速工业级AI仿真与自主系统落地。

  3. 14.
    从超级顾问到 Agent 蜂群：AI Native 人才服务组织的共生进化｜AICon深圳— InfoQ

    AI 导读 · AI原生组织从单点智能到群体协作的进化路径，揭示人才服务新模式。

  4. 15.
    ESP32-AI 亮相：8MB PSRAM 运行 2890 万参数模型，设想离线调配出最佳咖啡— IT之家

    AI 导读 · 低成本在低功耗芯片上跑大模型，让离线智能硬件成为可能。

  5. 16.
    聚灿光电：已对接多家AI眼镜企业推进定制研发— 36氪

    AI 导读 · 踩中AI眼镜风口，定制研发显示技术储备，但量产前需攻克客户端技术难题。

  6. 17.
    联想来酷斗战者公布战 7000P 锐龙版游戏本：8940HX + 5060— IT之家

    AI 导读 · 锐龙9配RTX 5060组合首次亮相，性能与性价比或成最大看点。

论文研究6 条

  1. 18.
    Certified Parallel-in-Time Sinkhorn for Dynamic Entropic Optimal Transport— arXiv

    AI 导读 · 并行时序Sinkhorn算法突破，加速动态最优传输计算，助力流匹配等前沿应用。

  2. 19.
    Learning Distributions from Multiple Data Providers— arXiv

    AI 导读 · 多数据源分布学习突破，解决异构重叠数据训练难题。

  3. 20.
    KANEx: Translating Kolmogorov-Arnold Networks' Interpretability to Medical Explainability— arXiv

    AI 导读 · 用KAN网络破解医学影像黑箱，可解释性突破让AI诊断更可信。

  4. 21.
    Global Convergence of DGM and PINN Algorithms for Solving Nonlinear PDEs— arXiv

    AI 导读 · 首次证明DGM与PINN求解非线性PDE的全局收敛性，为AI科学计算奠定理论基础。

  5. 22.
    The Physics of Multi-Turn Long-Horizon Planning: From Pre-training to Post-training via Single- and Multi-Teacher On-Policy Agentic Distillation— arXiv

    AI 导读 · 探索从预训练到后训练提升大模型长程规划能力，为自主智能体发展提供关键路径。

  6. 23.
    DataOrchestra: Learning to Orchestrate Per-Example Curation of Pretraining Data— arXiv

    AI 导读 · 按需定制每条训练数据，突破粗放式数据处理瓶颈。

快讯

日报生成时间：2026-07-28
← 2026-07-27最新日报没有更新 →