News Categories

Google’s AlphaGo triumphs over Go champion (Update)

By Chong Jinn Wei - on 16 Mar 2016, 10:20am

Google’s AlphaGo triumphs over Go champion (Update)

Update (March 16/2016): Google's AlphaGo A.I. has been declared the champion of the five-round Go game, having beaten the world's best Go player, Lee Sedol, who conceded defeat by resignation after 280 moves in the allotted two hours time, and two byō-yomi overtime periods.

The victory score for the AlphaGo A.I at the end was four-to-one, with Sedol himself putting up a tough fight against the A.I. in order to secure at least one victory against it.

With its victory, Google DeepMind will donate US$1 million (approx. RM4.13 million) to a multitude of charities, as well as several Go organizations.


Original article below:

Google DeepMind’s AlphaGo managed to triumph over Lee Sedol, the best Go player in the past decade. As of this writing, the machine had bested Lee twice in a best of five, marking the first time in history that an Artificial Intelligence (AI) managed to defeat a top-ranking Go player without any handicap.

Google DeepMind’s AlphaGo has managed to win 2-0 again Lee Sedol, Go Champion.

Previously, DeepMind, the A.I. branch of Google, made headway when it’s Go-playing computer program, AlphaGo, managed to beat Fan Hui, three-time European Go Champion. This isn't an easy feat, as Go is arguably the most complex game an AI can play, with over 200 possible moves for a given turn. This is why programmers see a Go playing AI to be a significant milestone in AI programming.

“What an incredibly exciting game. Lee Sedol is a formidable opponent, famed for his fighting style, and AlphaGo decided to go toe-to-toe with him, which made for a tense, close-fought game. We still have four games to go, so anything can still happen. Whatever the outcome, we feel this match is a testament to the power of human ingenuity,” said Dennis Hassabis, Co-Founder, CEO, DeepMind.

During the match, man versus machine were pretty evenly matched. Lee Sedol deployed aggressive tactics, while the AlphaGo AI answered back in kind. After 186 moves, AlphaGo was declared the victor after Lee Sedol conceded. In the second game, AlphaGo won again after 211 moves and Lee Sedol conceding a second time.

“I was surprised…If you look at the way the game was played, it was a very clear loss on my part. I felt like AlphaGo played certain problematic positions (the first match), but today (the second match) I felt that AlphaGo played a near perfect game. There was not a moment I felt like its moves were unreasonable,” said Lee Sedol, Go Champion.

Either AlphaGo or Lee Sedol must win at least three of the five games to determine who the true victor is. The next three games are set to go on Saturday March 12 (today), Sunday March 13 and Tuesday March 15 in Korea. The winner will receive USD$1million (~RM4,072,250), if AlphaGo wins the prize will be donated to Go organizations, UNICEF and STEM charities.

For the latest on Google, follow us here.