<?xml version="1.0" encoding="utf-8" standalone="yes"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
  <channel>
    <title>Information Retrieval on Nalar</title>
    <link>https://nalar.dev/tags/information-retrieval/</link>
    <description>Recent content in Information Retrieval on Nalar</description>
    <generator>Hugo</generator>
    <language>en-us</language>
    <lastBuildDate>Wed, 09 Sep 2026 00:00:00 +0700</lastBuildDate>
    <atom:link href="https://nalar.dev/tags/information-retrieval/index.xml" rel="self" type="application/rss+xml" />
    <item>
      <title>Diversify RAG Retrieval with Maximum Marginal Relevance</title>
      <link>https://nalar.dev/diversify-rag-retrieval-with-maximum-marginal-relevance/</link>
      <pubDate>Wed, 09 Sep 2026 00:00:00 +0700</pubDate>
      <guid>https://nalar.dev/diversify-rag-retrieval-with-maximum-marginal-relevance/</guid>
      <description>&lt;p&gt;A retrieval-augmented generation (RAG) system can retrieve highly relevant chunks and still build a poor context.&lt;/p&gt;&#xA;&lt;p&gt;The problem is redundancy. Imagine a support assistant answering a question about an API timeout. Vector search returns five chunks, but four are slightly different copies of the same timeout definition. The fifth useful chunk about retry behavior never reaches the model. Each result looked relevant in isolation, yet the set wastes most of its context budget repeating one idea.&lt;/p&gt;</description>
    </item>
  </channel>
</rss>
