AlphaGo

AlphaGo was developed by DeepMind, Google's London-based subsidiary, as part of a research project that began to take shape around 2014. Go is considered far more complex than chess. The number of possible moves is too vast to calculate ahead the way Deep Blue did. So AlphaGo combined two components: a "policy network" to identify promising moves and a "value network" to evaluate positions, together with a search method called Monte Carlo tree search. The system was first trained on about 30 million moves by human experts, and then improved by playing against itself (reinforcement learning).

In October 2015, AlphaGo beat European champion Fan Hui (a 2-dan professional) 5–0. It was the first time a computer program beat a professional Go player on a full board (19×19) without a handicap. The biggest win of all came in March 2016, when it faced Lee Sedol, a legendary South Korean 9-dan professional, in a five-game match in Seoul (9, 10, 12, 13 and 15 March). AlphaGo won 4–1. Lee won only the fourth game, with an extraordinary move 78, which the Chinese player Gu Li called a "divine move". According to DeepMind's Aja Huang, the move exposed a weakness in AlphaGo that was fixed in later versions. AlphaGo's own famous move 37 in the second game stunned even veteran Go experts with how far it departed from accepted human strategy.

In May 2017, the "Master" version of AlphaGo beat the world's top-ranked player, Ke Jie, 3–0 in Wuzhen, China. After that, DeepMind retired AlphaGo from official competition. The next version, AlphaGo Zero, was trained purely on self-play, with no human game data at all, and beat the original AlphaGo 100–0. Its successor, AlphaZero, extended the method to chess and shogi as well, with only 24 hours of training. The journal Science chose the 2016 achievement as a runner-up for "Breakthrough of the Year." Experts had expected such an achievement to take at least another 5–10 years of development. The win over Lee Sedol was broadcast live to tens of millions of viewers across Asia, and AlphaGo's achievement as a whole was described in China as a "Sputnik moment," an event that helped persuade the Chinese government to prioritize artificial intelligence and dramatically increase national funding for it.