Google’s AI AlphaGo has mastered the art of Go

Google’s London based Artificial Intelligence (AI) company, DeepMind developed AI software, AlphaGo which has mastered the art of playing Chinese board game, Go.

AlphaGo, combines an advanced tree search with deep neural networks. These neural networks take a description of the Go board as an input and process it through 12 different network layers containing millions of neuron-like connections. One neural network, the “policy network,” selects the next move to play. The other neural network, the “value network,” predicts the winner of the game.

AlphaGo learned to discover new strategies for itself, by playing thousands of games between its neural networks, and adjusting the connections using a trial-and-error process known as reinforcement learning.

“Many of the best programmers in the world were asked last year how long it would take for a program to beat a top professional, and most of them were predicting 10-plus years,” said DeepMind’s chief executive, Demis Hassabis.

“The reasons it was quicker than people expected was the pace of the innovation going on with the underlying algorithms and also how much more potential you can get by combining different algorithms together.”

Alpha Go was then put to test where first, it won all of its 500 games except one against other top programs at the forefront of computer Go. Second, in a closed doors match last October against three-time reigning European Go champion Fan Hui—an elite professional player who has devoted his life to Go since the age of 12 in the London office and it won by 5 games to 0. This was the first time a computer program had ever beaten a professional Go player.

In March, AlphaGo will face its ultimate challenge: a five-game challenge match in Seoul against the legendary Lee Sedol—the top Go player in the world over the past decade.