<?xml version="1.0" encoding="utf-8" standalone="yes"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
  <channel>
    <title>光学压缩 on AI内参</title>
    <link>https://www.neican.ai/tags/%E5%85%89%E5%AD%A6%E5%8E%8B%E7%BC%A9/</link>
    <description>Recent content in 光学压缩 on AI内参</description>
    <generator>Hugo</generator>
    <language>zh-cn</language>
    <lastBuildDate>Fri, 24 Oct 2025 20:10:09 +0800</lastBuildDate>
    <atom:link href="https://www.neican.ai/tags/%E5%85%89%E5%AD%A6%E5%8E%8B%E7%BC%A9/index.xml" rel="self" type="application/rss+xml" />
    <item>
      <title>DeepSeek又在整活儿了：AI“看图识字”还学会“选择性遗忘”，这波操作太秀！</title>
      <link>https://www.neican.ai/insights/deepseekai-20251024201009836-0/</link>
      <pubDate>Fri, 24 Oct 2025 20:10:09 +0800</pubDate>
      <guid>https://www.neican.ai/insights/deepseekai-20251024201009836-0/</guid>
      <description>DeepSeek团队再次“整活儿”，推出小模型DeepSeek-OCR，它通过“光学压缩”技术，让AI把长文本先变成图片再识别，巧妙地解决了大模型处理长文本的算力“瓶颈”，效率直接翻倍。更炸裂的是，这种视觉压缩还能模拟人类的“选择性遗忘”机制，给AI记忆管理提供了全新思路，简直是科技界的一股清流，用“小而美”颠覆了“大而全”的传统路径。</description>
    </item>
  </channel>
</rss>
