<?xml version="1.0" encoding="utf-8" standalone="yes"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
  <channel>
    <title>Consistency Regularization on Nalar</title>
    <link>https://nalar.dev/tags/consistency-regularization/</link>
    <description>Recent content in Consistency Regularization on Nalar</description>
    <generator>Hugo</generator>
    <language>en-us</language>
    <lastBuildDate>Fri, 04 Sep 2026 00:00:00 +0700</lastBuildDate>
    <atom:link href="https://nalar.dev/tags/consistency-regularization/index.xml" rel="self" type="application/rss+xml" />
    <item>
      <title>Learn from Unlabeled Data with Mean Teacher</title>
      <link>https://nalar.dev/learn-from-unlabeled-data-with-mean-teacher/</link>
      <pubDate>Fri, 04 Sep 2026 00:00:00 +0700</pubDate>
      <guid>https://nalar.dev/learn-from-unlabeled-data-with-mean-teacher/</guid>
      <description>&lt;p&gt;Many machine learning projects have far more raw examples than labeled ones. A team may have millions of images, audio clips, or sensor readings, but only a small subset has been reviewed by people. Standard supervised training ignores the unlabeled remainder because it has no target labels to compare with the model&amp;rsquo;s predictions.&lt;/p&gt;&#xA;&lt;p&gt;&lt;strong&gt;Mean Teacher&lt;/strong&gt; provides a way to use those unlabeled examples without pretending that their unknown labels are known. It trains a student model to make predictions that stay consistent with a more slowly changing teacher model. The teacher is not a separately trained expert: its parameters are an exponential moving average of the student&amp;rsquo;s parameters.&lt;/p&gt;</description>
    </item>
  </channel>
</rss>
