<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>WhiteningBERT: An Easy Unsupervised Sentence Embedding Approach · HackerPapers discussions</title><link>https://hackerpaper.kaipfister.de/p/2104.01767/</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/2104.01767/feed.xml" rel="self" type="application/rss+xml" /><item><title>WhiteningBERT: An Easy Unsupervised Sentence Embedding Approach — mentioned by cyanydeez</title><link>https://news.ycombinator.com/item?id=49513933</link><guid isPermaLink="false">urn:hackerpapers:mention:2104.01767:49513933</guid><pubDate>Mon, 31 Aug 2026 19:33:22 GMT</pubDate><description>&lt;p&gt;we want some kind of fractal knowledge graph. It starts coarse at some zoom level and you can move in and out. You can insert knowledge at any level and it&amp;#x27;s out/in levels adjust accordingly. The search semantics at every level are the same, but whats in the visible area changes based on what your focus is. One way I&amp;#x27;ve toyed with a graph outline with it using &amp;quot;whitening&amp;quot; (https://arxiv.org/pdf/2104.01767v3) for embeddings rather than just text, so you add things like file path, nearest title/method/const etc. You have to have dummy text though because it fails with &amp;quot;null&amp;quot;; all embeddings…&lt;/p&gt;&lt;p&gt;&lt;a href="https://hackerpaper.kaipfister.de/p/2104.01767/"&gt;Paper and discussion history&lt;/a&gt;. A mention is not an endorsement.&lt;/p&gt;</description><category>cs.CL</category></item></channel></rss>
