<?xml version="1.0" encoding="utf-8" standalone="yes"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:content="http://purl.org/rss/1.0/modules/content/">
  <channel>
    <title>Privacy on Bingjie&#39;s Blog</title>
    <link>https://blog.bj-yan.top/en/tags/privacy/</link>
    <description>Recent content in Privacy on Bingjie&#39;s Blog</description>
    <image>
      <url>https://blog.bj-yan.top/favicon.png</url>
      <link>https://blog.bj-yan.top/favicon.png</link>
    </image>
    <generator>Hugo -- gohugo.io</generator>
    <language>en-US</language>
    <lastBuildDate>Fri, 28 Jul 2023 14:55:49 +0800</lastBuildDate><atom:link href="https://blog.bj-yan.top/en/tags/privacy/index.xml" rel="self" type="application/rss+xml" />
    <item>
      <title>Information Theory in Privacy-Preserving Computation: How to Understand and Measure Information Leakage</title>
      <link>https://blog.bj-yan.top/en/p/blog-information-theory-for-privacy/</link>
      <pubDate>Fri, 28 Jul 2023 14:55:49 +0800</pubDate>
      
      <guid>https://blog.bj-yan.top/en/p/blog-information-theory-for-privacy/</guid>
      <description>Starting from the attacker&amp;#39;s observations and auxiliary information, this article introduces entropy, conditional mutual information, the data processing inequality, and Fano&amp;#39;s inequality. It also uses randomized response to illustrate the connections and distinctions between information-theoretic privacy and differential privacy.</description>
    </item>
    
    <item>
      <title>Privacy Leakage in Deep Learning: What Models, Predictions, and Gradients Reveal</title>
      <link>https://blog.bj-yan.top/en/p/blog-privacy-leakage-in-deep-learning/</link>
      <pubDate>Fri, 28 Jul 2023 09:54:50 +0800</pubDate>
      
      <guid>https://blog.bj-yan.top/en/p/blog-privacy-leakage-in-deep-learning/</guid>
      <description>Reviewing membership inference, attribute inference, model inversion, and training data extraction; deriving the recoverability of single-sample gradients; and discussing the applicability boundaries of federated learning, risk assessment, and various privacy-preserving techniques.</description>
    </item>
    
  </channel>
</rss>
