Balancing Cost and Effectiveness in Document Review Workflows

David Lewis, executive vice president for AI research, development and Ethics at Reveal-Brainspace will discuss manual reviews of large text collections and streams that increasingly leverage supervised learning, particularly active learning of text classifiers. Applications include electronic discovery (eDiscovery) in the law, systematic review in medicine, and content moderation of social media. These human-in-the-loop workflows must achieve high effectiveness while minimizing the cost of (and sometimes harm to) human reviewers. Lewis presents a new analytic framework and visualization tool for understanding these tradeoffs.  

 


 

Wednesday, November 3, 2021 at 1:45pm to 2:45pm

Center for Academic Success and Engagement (CASE), W262

Event Type

Lecture/Presentation

Interests

Communication & Information

Audience

Students, Faculty, Graduate Students

College, School & Unit

Media, Communication and Information

Group
CMCI: Media, Communication and Information
Add to Calendar
GoogleiCalOutlook

Recent Activity