<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>RadioTalk: a large-scale corpus of talk radio transcripts · HackerPapers discussions</title><link>https://hackerpaper.kaipfister.de/p/1907.07073/</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/1907.07073/feed.xml" rel="self" type="application/rss+xml" /><item><title>RadioTalk: a large-scale corpus of talk radio transcripts — mentioned by kick</title><link>https://news.ycombinator.com/item?id=22040285</link><guid isPermaLink="false">urn:hackerpapers:mention:1907.07073:22040285</guid><pubDate>Tue, 14 Jan 2020 00:10:43 GMT</pubDate><description>&lt;p&gt;Earlier this year the Media Lab did an absolutely ginormous automated transcription project. Off the top of my head, it was ~2.8 billion words. 13.1% error rate (vs. ~7% error rate for Google&amp;#x27;s proprietary solution). Found the paper on it: https://arxiv.org/pdf/1907.07073.pdf&lt;/p&gt;&lt;p&gt;&lt;a href="https://hackerpaper.kaipfister.de/p/1907.07073/"&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>
