🌳

How can a computer learn Go by playing random games?

Instead of calculating every possible move, Go programs learned to play out random games and count which moves tended to win. AlphaGo added neural networks and beat Lee Sedol.

Tap to go on, or swipe →