Learning in Real-Time in Repeated Games Using Experts (Extended Abstract)
Jacob W. Crandall · 2013
Despite much progress, state-of-the-art learning algorithms for repeated games still often require thousands of moves to learn eectively { even in simple games. Our goal is to nd algorithms that learn to play eective strategies in tens of moves in many games when paired against various associates. Toward this end, we describe a new meta-algorithm designed to increase the learning speed and prociency of expert algorithms. We show that this meta-algorithm enhances four expert algorithms so that they quickly learn effective strategies in two-player repeated games.