Move Evaluation in Go Using Deep Convolutional Neural Networks

TL;DR

This paper demonstrates that deep convolutional neural networks can effectively evaluate Go positions and moves, achieving performance comparable to expert players and surpassing traditional search-based programs.

Contribution

It introduces a deep convolutional neural network trained on professional games to evaluate Go positions, achieving high accuracy and strong gameplay performance without search.

Findings

Correctly predicts expert moves in 55% of positions

Outperforms GnuGo in 97% of games

Matches state-of-the-art Monte-Carlo tree search performance

Abstract

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.