<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Differentiable Ranks and Sorting using Optimal Transport · HackerPapers discussions</title><link>https://hackerpaper.kaipfister.de/p/1905.11885/</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/1905.11885/feed.xml" rel="self" type="application/rss+xml" /><item><title>Differentiable Ranks and Sorting using Optimal Transport — mentioned by marco_z</title><link>https://news.ycombinator.com/item?id=42638062</link><guid isPermaLink="false">urn:hackerpapers:mention:1905.11885:42638062</guid><pubDate>Wed, 08 Jan 2025 20:22:56 GMT</pubDate><description>&lt;p&gt;It&amp;#x27;s incredible how many algorithms can be made differentiable end-to-end. E.g. in https://arxiv.org/abs/1905.11885 &amp;quot;Differentiable ranks and sorting using optimal transport&amp;quot; they show how a relaxation of an array sorting procedure leads to differentiable rank statistics (quantiles etc.). The underlying theory is quite close to what I show in the blog post. In DETR they don&amp;#x27;t actually use a differentiable Hungarian algorithm, it&amp;#x27;s only used to score their neural predictions outside of the training loop IIUC.&lt;/p&gt;&lt;p&gt;&lt;a href="https://hackerpaper.kaipfister.de/p/1905.11885/"&gt;Paper and discussion history&lt;/a&gt;. A mention is not an endorsement.&lt;/p&gt;</description><category>cs.LG</category></item></channel></rss>
