Here we introduce a new approach to computer Go that uses ‘value networks’ to evaluate board positions and ‘policy networks’ to select moves. David Silver, et al., "Mastering the game of Go with deep neural networks and tree search" https:…