Monday, November 18, 2013

Title: Using User Behavior to Evaluate the Quality of Crowd-Generated Content
Speaker: Jeff Rzeszotarski (Human Computer Interaction Institute) 
Date: Wednesday, November 20
Time: 12-1pm
Room: GHC 6501 

Users create, contribute, and disseminate an astonishing amount of information online. Yet, not all of it is valuable. While existing approaches can readily identify obvious quality problems in simple content, judging the quality of creative, complex, or subjective work remains a major challenge. In this talk I will describe a novel approach for understanding and evaluating the quality of user-generated content. Rather than look at the final products, I propose examining the way a person works as they create them. I will discuss my work developing predictive models and novel visualizations for the behavior of crowdworkers on Mechanical Turk and volunteers on Wikipedia as they generate content, and explore future applications of this approach towards helping users across the web create better content before they even press “submit”.

Jeff Rzeszotarski (rez-oh-tar-ski) is a 4th year PhD student in the Human-Computer Interaction Institute at Carnegie Mellon University. Advised by Dr. Aniket Kittur, his research focuses on crowdsourcing and social computing, studying techniques that support groups of people generating and consuming content online. His work has received a best paper award from ACM UIST and he is the recipient of a Microsoft Research Fellowship. Jeff holds a BA from Carleton College.