Underunderstood

By Select Works

The Case of the Blah Blah Blahs

S3 E8

🔊 Play episode (36 min)

Direct episode link

💬 Share episode

Published December 08, 2020 5:59am

A famous datset of Reuters articles from the 1980s includes “Blah blah blah.” in place of some stories. Why?

We have a Patreon now! Sign up to support the show and get access to our bonus podcast, Overunderstood.

Show notes: 00:31 - The link Jess sent 8:31 - SGML 8:46 - This is what the blahs look like and this is what all the entries look like. 24:00 - FTP 24:34 - Linguistic Data Consortium 29:00 - RCV1 at NIST and David D. Lewis’s README 30:22 - Construe-TIS: A System for Content-Based Indexing of a Database of News Stories (Phil Hayes...

Return to podcast