The facebook files at training time: incentivizing discrimination in machine learning

Location:

BBB3725/ Zoom (Zoom link; Password if needed: MichiganAI)

Abstract:

Recent work has documented instances of unfairness in deployed machine learning models, and significant research effort has been dedicated to creating algorithms that seek to ensure fairness. In this talk, I will highlight a motivating case study due to recent reporting by the Wall Street Journal on Facebook.  This case study demonstrates the need for a new approach to understanding and creating algorithms that consider fairness as a core value, one that considers the incentives and market forces that drive the creation of decision-making systems that employ machine learning.

This talk focuses in particular on the market forces that drive differential investment in the data pipeline for different groups.  We develop a high-level model to study this question. We show that our model predicts unfairness in a monopoly setting, and competition does not necessarily eliminate this tendency, and may even exacerbate it.   We also consider ways to regulate a machine-learning driven monopolist and quantify the “price” of fairness (and who pays it). This work implies that mitigating fairness concerns may require policy-driven solutions, not only technological ones.

Bio:  Currently Ben Fish is Assistant Professor at U-M CSE. Previously he was a postdoctoral fellow at Mila hosted by Fernando Diaz, which he joined after moving from the Fairness, Accountability, Transparency, and Ethics (FATE) Group at Microsoft Research Montréal, also hosted by Fernando Diaz.  His research develops methods for machine learning and other computational systems that incorporate human values and social context.  This includes scholarship in fairness and ethics in machine learning and learning over social networks.  He received his Ph.D. from the University of Illinois at Chicago as a member of the Mathematical Computer Science group.  He was previously a visiting researcher at the University of Melbourne and the University of Utah, and earned a B.A. from Pomona College in Mathematics and Computer Science.