Humans have come to rely on machines to make decisions concerning their welfare. When these learning based decision-making systems interact with humans, they present interesting challenges. For instance, when the decision rule is known, rational agents may respond to it by manipulating their features to obtain favorable outcomes. In such a case, the goal is to find the most robust strategy decision rule. In this talk, I will present two complementary frameworks; strategic classification and strategic representation. In strategic classification, agents misrepresent their features to game the system's classifier to gain a favorable outcome. Tables are turned in the strategic representation setting where the system (recommendation system, for instance), strategically curates the information presented to the agents/decision-makers to induce favorable decisions. I will present a few strategies robust algorithms in these settings and discuss their limitations. Part of the work presented in this talk is in collaboration with Vineet Nair, Nir Rosenfeld, Inbal Talgam-Cohen and Itay Eilat.
Dr. Ganesh Ghalme is an assistant professor at the Department of AI, IIT Hyderabad. He was a postdoctoral fellow at the Game theory group, Technion-Israel Institute of Technology, Israel from July 2020 to Mar 2022. Prior to that, he completed his PhD from IISc Bangalore in June 2020. His research interests lie at the intersection of game theory and machine learning. In particular, he works in fairness in online learning, strategic learning, information design and fair division.