AlphaGo wrapped up victory for Google in the DeepMind Challenge Match by winning its third straight game against Go champion Lee Se-dol yesterday, but the 33-year-old South Korean has got at least some level of revenge — he's just defeated AlphaGo, the AI program developed by Google's DeepMind unit, in the fourth game of a five-game match in Seoul.
AlphaGo is now 3-1 up in the series with a professional record, if you can call it that, of 9-1 including the 5-0 win against European champion Fan Hui last year. Lee's first win came after an engrossing game where AlphaGo played some baffling moves, prompting commentators to wonder whether they were mistakes or — as we've often seen this week — just unusual strategies that would come good in the end despite the inscrutable approach. (To humans, at least.)
Lee Sedol is playing brilliantly! #AlphaGo thought it was doing well, but got confused on move 87. We are in trouble now...— Demis Hassabis (@demishassabis) March 13, 2016
According to tweets from DeepMind founder Demis Hassabis, however, this time AlphaGo really did make mistakes. The AI "thought it was doing well, but got confused on move 87," Hassabis said, later clarifying that it made a mistake on move 79 but only realized its error by 87. AlphaGo adjusts its playing style based on its evaluation of how the game is progressing.
Lee entered the post-game press conference to rapturous applause, remarking "I've never been congratulated so much just because I won one game!" Lee referred back to his post-match prediction that he would win the series 5-0 or 4-1, saying that this one win feels even more valuable after losing the first three games.
"Lee Se-dol is an incredible player and he was too strong for AlphaGo today," said Hassabis, adding that the defeat would help DeepMind test the limits of its AI. "For us this loss is very valuable. We're not sure what happened yet."
DeepMind's AlphaGo program has beaten 18-time world champion Lee three times so far with its advanced system based on deep neural networks and machine learning. The series is the first time a computer program has taken on a professional 9-dan player of Go, the ancient Chinese board game long considered impossible for computers to play at a world-class level due to the high level of intuition required to master its intricate strategies. Lee was competing for a $1 million prize put up by Google, but DeepMind's victory means the sum will be donated to charity.