AlphaGo, by Google DeepMind. It beat European champion Fan Hui 5-0 in October 2015, the first win by a computer over a professional on a full 19 × 19 board without handicap.
Card 1 of 259. Answer: AlphaGo, by Google DeepMind. It beat European champion Fan Hui 5-0 in October 2015, the first win by a computer over a professional on a full 19 × 19 board without handicap.
About · Quizzes · History of AI · To suggest new questions or report an error send me a dm.
AlphaGo, a program built by Google DeepMind. In October 2015 it beat Fan Hui, the reigning European champion and a 2-dan professional, 5 games to 0 in London. It was the first time a computer had beaten a professional Go player in an even game, with no handicap stones, on the full 19 × 19 board.
DeepMind announced the result on 27 January 2016 with a Nature paper, "Mastering the game of Go with deep neural networks and tree search". Go had resisted the brute-force search that beat chess, with about 250 legal moves per turn against roughly 35. Earlier programs beat professionals only with a head start: Crazy Stone beat the 9-dan Ishida Yoshio in 2013 with four handicap stones.
AlphaGo paired Monte Carlo tree search with two networks: a policy network that proposes moves, trained on 30 million positions from expert games, and a value network that judges positions, trained on self-play games. The better-known 4 to 1 win over Lee Sedol, a 9-dan, came later, in March 2016.
