Monday / CS Seminar

November 20th / 12:00 PM

Hybrid / Mudd 3514

Speaker: Doug Downey

Title: High-precision NLP at discount prices: Lessons learned in Semantic Scholar

Abstract

Natural language processing (NLP) has made major strides in recent years, due to the increasing capabilities of large language models. However, using NLP to power real applications is still challenging: the best models are expensive to use at scale and are still prone to errors. I'll describe recent lessons we've learned on the Semantic Scholar team as we've built and deployed applications using NLP aimed at accelerating science, including PDF content extraction, an AI-powered reading interface, and automatically-constructed topics pages for science. While recent NLP breakthroughs do enable exciting new experiences, more long-standing considerations from applied machine learning---like task selection and model design---remain critical.

Biography

Doug Downey is professor of Computer Science at Northwestern University, and Senior Director of Semantic Scholar Research at the Allen Institute for AI (where he currently serves full-time). The Semantic Scholar Research group performs NLP and HCI research aimed at helping scholars discover, understand, and synthesize scientific literature.

Zoom Link: https://northwestern.zoom.us/j/93420493721?pwd=YzgxcFVaQUxiTDcySUd2WEtGTG94Zz09

Panopto: https://northwestern.hosted.panopto.com/Panopto/Pages/Viewer.aspx?id=ca48eade-a897-47c4-8d47-b0bc0145a676

Research Interests/Area

Natural language processing, human-computer interaction