Learning Algorithms and Society: Case Studies
Bo Waggoner - Postdoctoral Fellow, University of Pennsylvania, Warren Center for Network and Data Sciences
Jan. 18, 2018, 10 a.m. - Jan. 18, 2018, 11:30 a.m.
ENGMC 321
This talk will consider algorithms for learning or making decisions based on data -- in a societal context. The goal is to understand, and design, algorithms that affect people in terms of (1) fairness and (2) privacy. First, we'll see a model for learning and decision-making over time when decisions may affect people in an unfair or discriminative way. Second, we'll investigate machine learning when data comes from people who have privacy concerns, and ways to transform theoretical results into practical solutions. In both cases, we'll discuss how to formalize social concepts mathematically, and consider how social constraints change design or performance of traditional algorithms.
Based on joint works with Sampath Kannan, Katrina Ligett, Jamie Morgenstern, Seth Neel, Aaron Roth, and Steven Wu.
Bio: Bo Waggoner is a postdoctoral fellow at the University of Pennsylvania's Warren Center for Network and Data Sciences. His work focuses on systems for learning and aggregating information in contexts with strategic behavior, privacy, or fairness considerations. He received his PhD from Harvard in 2016.
Everyone is welcome
