Tuesday 18 June 2018 16:00 – 17:00 A. Payatakes Seminar Room
“From Games to Algorithms, Telecommunications and Machine Learning”
Dr. Fasoulakis Michail Institute of Computer Science (ICS)
Abstract
Game theory is the area of mathematics for analysing conflict and cooperation situations. One of the most reasonable solution concepts in game theory is the Nash equilibrium, a stable solution that no player has any incentive to deviate. Nash equilibria could a priori predict what rational players should do in a competitive situation. But what is the prediction power if we cannot compute them? The problem of computing a Nash equilibrium is now known to be complete in the computational complexity class PPAD, that makes the area of computing approximate Nash equilibria one of the most important areas in equilibrium computation. We will present the state-of-the-art in computation of approximate Nash equilibria. At the second part of the presentation, we will show applications of (algorithmic) game theory in the area of telecommunications, for instance in a museum environment. Furthermore, we will present a game theoretic study of jamming in multiple Gaussian independent channels and finally, we will present the Generative Adversarial Networks and the formulation of them as zero-sum game in the intersection of game theory and machine learning.