<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Mask-Predict: Parallel Decoding of Conditional Masked Language Models · HackerPapers discussions</title><link>https://hackerpaper.kaipfister.de/p/1904.09324/</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/1904.09324/feed.xml" rel="self" type="application/rss+xml" /><item><title>Mask-Predict: Parallel Decoding of Conditional Masked Language Models — mentioned by jaaustin</title><link>https://news.ycombinator.com/item?id=45644944</link><guid isPermaLink="false">urn:hackerpapers:mention:1904.09324:45644944</guid><pubDate>Mon, 20 Oct 2025 15:19:18 GMT</pubDate><description>&lt;p&gt;To my knowledge this connection was first noted in 2021 in https://arxiv.org/abs/2107.03006 (page 5). We wanted to do text diffusion where you’d corrupt words to semantically similar words (like “quick brown fox” -&amp;gt; “speedy black dog”) but kept finding that masking was easier for the model to uncover. Historically this goes back even further to https://arxiv.org/abs/1904.09324, which made a generative MLM without framing it in diffusion math.&lt;/p&gt;&lt;p&gt;&lt;a href="https://hackerpaper.kaipfister.de/p/1904.09324/"&gt;Paper and discussion history&lt;/a&gt;. A mention is not an endorsement.&lt;/p&gt;</description><category>cs.CL</category></item><item><title>Mask-Predict: Parallel Decoding of Conditional Masked Language Models — HN submission</title><link>https://news.ycombinator.com/item?id=20886448</link><guid isPermaLink="false">urn:hackerpapers:submission:1904.09324:20886448</guid><pubDate>Thu, 05 Sep 2019 13:30:18 GMT</pubDate><description>&lt;p&gt;Mask-Predict: Parallel Decoding of Conditional Masked Language Models&lt;/p&gt;&lt;p&gt;&lt;/p&gt;&lt;p&gt;1 point; 0 discussion replies on HN. &lt;a href="https://hackerpaper.kaipfister.de/p/1904.09324/"&gt;Paper and source history&lt;/a&gt;. A submission is not an endorsement.&lt;/p&gt;</description><category>cs.CL</category></item></channel></rss>
