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.

Keywords

Monte Carlo tree searchComputer scienceConvolutional neural networkArtificial intelligenceDeep learningMachine learningMonte Carlo method

Affiliated Institutions

Related Publications

GameFlow

Although player enjoyment is central to computer games, there is currently no accepted model of player enjoyment in games. There are many heuristics in the literature, based on ...

2005 Computers in entertainment 2008 citations

Publication Info

Year
2014
Type
article
Citations
92
Access
Closed

External Links

Social Impact

Social media, news, blog, policy document mentions

Citation Metrics

92
OpenAlex

Cite This

Chris J. Maddison, Aja Huang, Ilya Sutskever et al. (2014). Move Evaluation in Go Using Deep Convolutional Neural Networks. arXiv (Cornell University) . https://doi.org/10.48550/arxiv.1412.6564

Identifiers

DOI
10.48550/arxiv.1412.6564