Abstract
We study the feedback loops between ML algorithms and strategic human subjects, and their implications for the design of (fair) ML. To this end, we analyze a strategic classification problem in which agents can adjust their behavior in response to a (fair) classifier deployed by the firm. Agents can opt to “manipulate” the algorithm (by only changing their observable features without changing their unobservable qualification state), or invest effort in “improvement” (which also changes their qualification states). We discuss the potential equilibria when these two options have different costs and efficacy, as well as how the firm can leverage this understanding to design incentive mechanisms which encourage improvements while discouraging manipulations.
Biography
Parinaz Naghizadeh is an assistant professor in the Integrated Systems Engineering and Electrical and Computer Engineering departments at The Ohio State University. She received her PhD in electrical engineering from the University of Michigan in 2016. Her research interests are in network economics, game theory, algorithmic economics, and reinforcement learning. She is a recipient of the NSF CAREER award in 2022, a Rising Stars in EECS in 2017, and a Barbour Scholarship in 2014.
