In this talk I will give an overview of my research interests and I will present our recent contributions in the area of computation of equilibria. Namely, I will present two polynomial-time algorithms for computing (1/3+δ)-Nash equilibria and (1/2+δ)-well-supported Nash equilibria in bimatrix games, for δ>0, two problems open for 15 and 7 years from the previous bounds, respectively and a new learning algorithm for achieving last iterate convergence to a Nash equilibrium in the class of zero-sum bimatrix games. Finally, I will give new future directions in this area.
Dr. Michail Fasoulakis is an (Applied) Mathematician and (Theoretical) Computer Scientist, with a background and interest in Computer Engineering and (Computational) Economics/Game theory/Operations Research. He holds a BSc in Computer Science, a second BSc in Mathematics, and a MSc in Communications and Signal Processing, all from the University of Crete. Furthermore, he holds a MSc in Computation and Game theory from the University of Liverpool, UK. He received the PhD in (theoretical) Computer Science from the University of Warwick UK, in 2017, working in the research area of algorithmic game theory, where he was a member of the Centre for Discrete Mathematics and its Applications.
After his PhD studies he had research fellow/postdoc positions in various academic places such as Athens University of Economics and Business (AUEB), Center for Mathematics and Computer Science in Netherlands - Centrum Wiskunde & Informatica/CWI (as an ERCIM "Alain Bensoussan" fellow), Foundation for Research and Technology–Hellas (FORTH) as a Stavros Niarchos Foundation (SNF)-FORTH fellow for a period, School of Electrical and Computer Engineering of the National Technical University of Athens, and University of Crete. He is currently an Affiliated Researcher with the Institute of Computer Science of FORTH and the Department of Informatics of AUEB.
His research interests include the (intersections of) areas of Algorithms, Game theory, Optimization, Mathematics of Information (Information theory), and their (theoretical) applications, especially in Operations Research, Economics, AI/ML, Communications, and Signal Processing. His work has been published in high quality journals and conferences' proceedings such as ACM Transactions on Algorithms, Algorithmica, IEEE Transactions on Neural Networks and Learning Systems, SIAM Journal on Computing, AAAI, AISTATS, ESA, and SODA, with notable contributions, with his coauthors, the current state of art polynomial-time algorithms for computing approximate (well-supported) Nash equilibria in bimatrix games, giving the best approximation bounds at the moment of the publications. During his studies and career, he has been awarded with a number of research grants/fellowships-scholarships/awards.