<?xml version="1.0" encoding="utf-8" standalone="yes"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
  <channel>
    <title>学术范式 on AI内参</title>
    <link>https://www.neican.ai/tags/%E5%AD%A6%E6%9C%AF%E8%8C%83%E5%BC%8F/</link>
    <description>Recent content in 学术范式 on AI内参</description>
    <generator>Hugo</generator>
    <language>zh-cn</language>
    <lastBuildDate>Thu, 16 Jul 2026 15:40:07 +0800</lastBuildDate>
    <atom:link href="https://www.neican.ai/tags/%E5%AD%A6%E6%9C%AF%E8%8C%83%E5%BC%8F/index.xml" rel="self" type="application/rss+xml" />
    <item>
      <title>重构学术话语权：WAICA如何以“AI原生”范式定义科研的未来</title>
      <link>https://www.neican.ai/insights/waicaai-20260716154007634-0/</link>
      <pubDate>Thu, 16 Jul 2026 15:40:07 +0800</pubDate>
      <guid>https://www.neican.ai/insights/waicaai-20260716154007634-0/</guid>
      <description>WAICA通过“AI原生”评审与展示范式，打破了传统学术会议的机制陈规，将科研成果从静态文本升级为可验证的智能资产，构建起连接学术研究与产业应用的闭环生态。</description>
    </item>
    <item>
      <title>从“技术膨胀”到“基建沉淀”：ICML 2026 揭示AI研究的范式转向</title>
      <link>https://www.neican.ai/insights/icml-2026-ai-20260706104007262-2/</link>
      <pubDate>Mon, 06 Jul 2026 10:40:07 +0800</pubDate>
      <guid>https://www.neican.ai/insights/icml-2026-ai-20260706104007262-2/</guid>
      <description>ICML 2026标志着AI科研从单纯规模竞赛转向对技术深度与伦理治理的深刻反思。会议通过表彰扩散模型底层优化及对对齐机制的批判性立场，预示了未来AI领域将向更高效、更透明及更具社会责任感的方向迈进。</description>
    </item>
    <item>
      <title>ScholAI：AI如何重塑学术研究的范式与前沿探索</title>
      <link>https://www.neican.ai/insights/scholaiai-20250624151004540-4/</link>
      <pubDate>Tue, 24 Jun 2025 15:10:04 +0800</pubDate>
      <guid>https://www.neican.ai/insights/scholaiai-20250624151004540-4/</guid>
      <description>ScholAI，一款基于多云平台（MCP）的智能学术研究工具，通过语义查询、多源检索和CCF排名集成等功能，显著提升了文献查找、分析和管理效率。该工具不仅预示着AI在科学发现领域的深入应用，重塑了传统学术研究范式，也引发了对未来研究工作效率提升与潜在伦理挑战的深层思考。</description>
    </item>
    <item>
      <title>超越顶会：一篇博客文章如何颠覆AI研究的价值衡量</title>
      <link>https://www.neican.ai/insights/article-20250617003004870-2/</link>
      <pubDate>Tue, 17 Jun 2025 00:30:04 +0800</pubDate>
      <guid>https://www.neican.ai/insights/article-20250617003004870-2/</guid>
      <description>一篇未经同行评审的博客文章，介绍了一款名为Muon的神经网络优化器，因其卓越的实测效果，助作者凯勒·乔丹获得OpenAI的Offer，并引发了AI研究领域对传统学术发表模式的深刻反思。文章深入分析了Muon的技术原理、其在效率上超越AdamW的表现，以及由月之暗面团队对Muon的进一步验证和开源，指出AI研究正从“论文至上”转向“实效为先”的范式转变。</description>
    </item>
    <item>
      <title>一篇博客直通OpenAI：深度学习优化器Muon如何重塑AI研究范式</title>
      <link>https://www.neican.ai/insights/openaimuonai-20250616163004/</link>
      <pubDate>Mon, 16 Jun 2025 16:30:04 +0800</pubDate>
      <guid>https://www.neican.ai/insights/openaimuonai-20250616163004/</guid>
      <description>凭借一篇关于新型深度学习优化器Muon的博客文章，而非传统学术论文，成功加入OpenAI，其技术或已用于GPT-5训练。这一事件不仅展示了Muon在提升AI模型训练效率方面的巨大潜力，更深层地揭示了AI研究领域正在经历的范式转变：即从传统的学术出版模式转向开放、社区驱动和以实际影响力为核心的创新与人才评估体系。</description>
    </item>
  </channel>
</rss>
