<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Diverse Beam Search: Decoding Diverse Solutions from Neural Sequence Models · HackerPapers discussions</title><link>https://hackerpaper.kaipfister.de/p/1610.02424/</link><description>Collected HN comments and submissions linking this paper.</description><language>en</language><lastBuildDate>Mon, 02 Nov 2026 11:16:22 GMT</lastBuildDate><atom:link href="https://hackerpaper.kaipfister.de/p/1610.02424/feed.xml" rel="self" type="application/rss+xml" /><item><title>Diverse Beam Search: Decoding Diverse Solutions from Neural Sequence Models — mentioned by eugenhotaj</title><link>https://news.ycombinator.com/item?id=21335247</link><guid isPermaLink="false">urn:hackerpapers:mention:1610.02424:21335247</guid><pubDate>Wed, 23 Oct 2019 14:56:22 GMT</pubDate><description>&lt;p&gt;I&amp;#x27;ve seen some research [1] where the authors use beam search with an explicit diversity penalty to get around the repetition problem. They seem to get good results. [1] https://arxiv.org/pdf/1610.02424.pdf&lt;/p&gt;&lt;p&gt;&lt;a href="https://hackerpaper.kaipfister.de/p/1610.02424/"&gt;Paper and discussion history&lt;/a&gt;. A mention is not an endorsement.&lt;/p&gt;</description><category>cs.AI</category></item></channel></rss>
