The game of Go is more challenging than other board games, due to the difficulty of constructing a position or move evaluation function. In this paper we investigate whether deep convolutional networks can be used to directly represent and learn this knowledge. We train a large 12-layer convolutional neural network by supervised learning from a database of human professional games. The network correctly predicts the expert move in 55% of positions, equalling the accuracy of a 6 dan human player. When the trained convolutional network was used directly to play games of Go, without any search, it beat the traditional search program GnuGo in 97% of games, and matched the performance of a state-of-the-art Monte-Carlo tree search that simulates a million positions per move.

—

Speaker’s Bio

Chris Maddison is a PhD student at the University of Toronto supervised by Geoff Hinton. His interests are primarily in inference and Monte Carlo methods. His work in this area has focused on generalizing techniques for sampling in discrete spaces to more general spaces and understanding annealing-based MCMC methods.