<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Combining Modular Skills in Multitask Learning · HackerPapers discussions</title><link>https://hackerpaper.kaipfister.de/p/2202.13914/</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/2202.13914/feed.xml" rel="self" type="application/rss+xml" /><item><title>Combining Modular Skills in Multitask Learning — mentioned by edwardjhu</title><link>https://news.ycombinator.com/item?id=35296885</link><guid isPermaLink="false">urn:hackerpapers:mention:2202.13914:35296885</guid><pubDate>Fri, 24 Mar 2023 23:08:56 GMT</pubDate><description>&lt;p&gt;This paper tries something like that https://arxiv.org/pdf/2202.13914.pdf The gain isn&amp;#x27;t that significant. We don&amp;#x27;t understand what these low-rank updates represent, and they might not correspond to &amp;quot;skills&amp;quot; that humans have.&lt;/p&gt;&lt;p&gt;&lt;a href="https://hackerpaper.kaipfister.de/p/2202.13914/"&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>
