ABSTRACT

Title of dissertation: COMPUTATIONAL ANALYSIS OF INTELLIGENT AGENTS: SOCIAL AND STRATEGIC SETTINGS

Anshul Sawant, Doctor of Philosophy, 2016

Dissertation co-directed by: Prof. V.S. Subrahmanian, Department of Computer Science Prof. Mohammad Taghi. Hajiaghayi, Department of Computer Science

The central motif of this work is prediction and optimization in presence of mul- tiple interacting intelligent agents. We use the phrase ‘intelligent agents’ to imply in some sense, a ‘’, the exact meaning of which varies depending on the setting. Our agents may not be ‘rational’ in the classical game theoretic sense, in that they don’t always optimize a global objective. Rather, they rely on heuristics, as is natural for human agents or even software agents operating in the real-world. Within this broad framework we study the problem of influence maximization in social net- works where behavior of agents is myopic, but complication stems from the structure of interaction networks. In this setting, we generalize two well-known models and give new algorithms and hardness results for our models. Then we move on to models where the agents reason strategically but are faced with considerable uncertainty. For such games, we give a new and analyze a real-world game using out techniques. Finally, the richest model we consider is that of Network Cournot Com- petition which deals with strategic resource allocation in hypergraphs, where agents reason strategically and their interaction is specified indirectly via player’s utility func- tions. For this model, we give the first equilibrium computability results. In all of the above problems, we assume that payoffs for the agents are known. However, for real- world games, getting the payoffs can be quite challenging. To this end, we also study the inverse problem of inferring payoffs, given game history. We propose and evaluate a data analytic framework and we show that it is fast and performant. COMPUTATIONAL ANALYSIS OF INTELLIGENT AGENTS: SOCIAL AND STRATEGIC SETTINGS

by

Anshul Sawant

Dissertation submitted to the Faculty of the Graduate School of the University of Maryland, College Park, in partial fulfillment of the requirements for the degree of Doctor of Philosophy 2016

Advisory Committee: Dr. V. S. Subrahmanian, Co-chair/Co-advisor Dr. MohammadTaghi Hajiaghayi, Co-chair/Co-advisor Dr. Sarit Kraus, Dept. of Computer Science, Bar-Ilan University Dr. Dana Nau Dr. Subramanian Raghavan, Robert H. Smith School of Business and Institute for Sys- tems Research © Copyright by Anshul Sawant 2016 Dedication

Dedicated to my mother, Dr. Satya Sawant.

ii Acknowledgments

To start with, I would like to thank my advisors, Prof. V.S. Subrahmanian and Prof. MohammadTaghi Hajiaghayi. Both have provided excellent guidance to me over the years. They have provided me with invaluable direction in my efforts and have been a constant source of ideas and inspiration. I would also like to thank Prof. Naveen Garg without whose guidance and inspiration, I would never have started this Ph.D. program. Over the course of my Ph.D. I have had the opportunity to collaborate with several wonderful people. I am thankful for contributions of (in no particular order) Hamid Mahini, David Malec, Sarit Kraus, Chanhyun Kang, Melika Abolhassani, John P. Dick- erson, Eric Demaine and S. Raghavan. I would like to thank Sarit Kraus, S.Raghavan and Dana Nau for serving on my dissertation committee. I am very thankful to Jennifer Story and Fatima Bangura. They made this such a smooth sailing. Never once though the course of my Ph.D did I have to think about bureaucratic hassles. Of course, none of this would have been any fun without friends. I would specif- ically like to thank Amit, Bhaskar, Kartik, Manish, Ramakrishna (guys at 5002), Meethu, Sudha (girls at 117), Ioana, Chanhyun, Eunhui, Noseong, Srijan