<?xml version="1.0" encoding="utf-8" standalone="yes"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
  <channel>
    <title>ICLR 2026 on AI内参</title>
    <link>https://www.neican.ai/tags/iclr-2026/</link>
    <description>Recent content in ICLR 2026 on AI内参</description>
    <generator>Hugo</generator>
    <language>zh-cn</language>
    <lastBuildDate>Tue, 12 May 2026 08:40:03 +0800</lastBuildDate>
    <atom:link href="https://www.neican.ai/tags/iclr-2026/index.xml" rel="self" type="application/rss+xml" />
    <item>
      <title>超越工程狂热：从 ICLR 2026 看中国 AI 的“务实范式”与原创性阵痛</title>
      <link>https://www.neican.ai/insights/-iclr-2026-ai--20260512084003106-4/</link>
      <pubDate>Tue, 12 May 2026 08:40:03 +0800</pubDate>
      <guid>https://www.neican.ai/insights/-iclr-2026-ai--20260512084003106-4/</guid>
      <description>ICLR 2026 的数据透视出中国 AI 研究正处于工程化执行力的巅峰，其低组织摩擦的协作模式构建了极强的技术追赶优势。然而，要实现从“规模化领先”到“范式定义者”的跨越，中国 AI 必须在下一阶段攻克原始创新能力的瓶颈。</description>
    </item>
    <item>
      <title>离了大谱！ICLR 2026 审稿竟有21%是AI“代笔”，学术圈也“塌房”了？</title>
      <link>https://www.neican.ai/insights/iclr-2026-21ai-20251117144006993-0/</link>
      <pubDate>Mon, 17 Nov 2025 14:40:06 +0800</pubDate>
      <guid>https://www.neican.ai/insights/iclr-2026-21ai-20251117144006993-0/</guid>
      <description>ICLR 2026顶会爆出惊天大瓜：第三方机构潘格拉姆实验室统计发现，竟有21%的审稿意见是AI纯度100%的“代笔”，而人类亲笔只占43%。这一现象引发社区热议，官方已紧急回应，表示正在讨论对策，并呼吁作者发现问题及时反馈，再次凸显了AI时代学术诚信面临的严峻挑战。</description>
    </item>
    <item>
      <title>ICLR 2026：AI写论文？不报就拒！顶会玩真的了，学术圈“瑟瑟发抖”？</title>
      <link>https://www.neican.ai/insights/iclr-2026ai-20250829114004619-1/</link>
      <pubDate>Fri, 29 Aug 2025 11:40:04 +0800</pubDate>
      <guid>https://www.neican.ai/insights/iclr-2026ai-20250829114004619-1/</guid>
      <description>ICLR 2026发布了“史上最严”大模型使用新规，要求论文作者和审稿人必须明确声明使用LLM，并对所有内容负全责，否则直接拒稿。此举旨在打击“提示词注入”等学术不端行为，确保学术诚信，尽管ICLR自身研究也显示AI在审稿中具有积极潜力，预示着AI与学术的“斗法”将走向更规范的共存模式。</description>
    </item>
  </channel>
</rss>
