
  <rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
    <channel>
      <title>d.eun.log</title>
      <link>https://deunlog.vercel.app/blog</link>
      <description>thoughts while i learn — code, tools, and everything in between.</description>
      <language>ko-KR</language>
      <managingEditor>daeunlee.dev@gmail.com (이다은)</managingEditor>
      <webMaster>daeunlee.dev@gmail.com (이다은)</webMaster>
      <lastBuildDate>Fri, 22 May 2026 00:00:00 GMT</lastBuildDate>
      <atom:link href="https://deunlog.vercel.app/feed.xml" rel="self" type="application/rss+xml"/>
      
  <item>
    <guid>https://deunlog.vercel.app/blog/ml-loss-gradient</guid>
    <title>모델이 틀렸을 때 어떻게 고치는가 — 손실 함수, 소프트맥스, 경사하강법</title>
    <link>https://deunlog.vercel.app/blog/ml-loss-gradient</link>
    <description>예측이 틀렸을 때 AI는 어떻게 스스로를 수정할까요. 손실 함수로 채점하고, 경사하강법으로 방향을 잡고, 소프트맥스로 확률을 만드는 과정을 정리했습니다.</description>
    <pubDate>Fri, 22 May 2026 00:00:00 GMT</pubDate>
    <author>daeunlee.dev@gmail.com (이다은)</author>
    <category>AI</category><category>머신러닝</category><category>딥러닝</category>
  </item>

  <item>
    <guid>https://deunlog.vercel.app/blog/ml-how-ai-learns</guid>
    <title>AI는 어떻게 배우는가 — 머신러닝 학습 방법 4가지</title>
    <link>https://deunlog.vercel.app/blog/ml-how-ai-learns</link>
    <description>지도학습, 비지도학습, 자기지도학습, 강화학습까지 — AI가 데이터에서 패턴을 찾는 4가지 방식과, 모델이 너무 잘 외워버렸을 때 생기는 문제까지 정리했습니다.</description>
    <pubDate>Sat, 02 May 2026 00:00:00 GMT</pubDate>
    <author>daeunlee.dev@gmail.com (이다은)</author>
    <category>AI</category><category>머신러닝</category><category>딥러닝</category>
  </item>

    </channel>
  </rss>
