This story is over 5 years old.


Google’s AI Is Now Reigning Go Champion of the World

It’s a pretty big moment for artificial intelligence.

On Saturday afternoon in Seoul, AlphaGo, the Go-playing artificial intelligence created by Google's DeepMind, beat 18-time Go world champion Lee Sedol for its third straight win in a five game series.

The win was a historic one for artificial intelligence research, a field where AI's mastery of this 2,500 year old game was long considered a holy grail of sorts for AI researchers. This win was particularly notable because the match included situations called ko fights which hadn't arisen in the previous two games. Prior to AlphaGo's win, other Go experts had speculated that ko situations could prove to be stumbling blocks for the DeepMind program as they had been in the past for other Go computer programs.


"When you watch really great Go players play, it is like a thing of beauty," said Google co-founder Sergey Brin, himself a self-proclaimed adamant Go player in grad school, after the match. "So I'm very excited that we've been able to instill that level of beauty inside a computer. I'm really honored to be here in the company of Lee Sedol, such an incredible player, as well as the DeepMind team who've been working so hard on the beauty of a computer."

AlphaGo is now poised for a repeat of its previous 5-0 sweep against 3-time European Go champion Fan Hui in October 2015, when it became the first computer program to beat a human Go player in an even match. Prior to Hui's loss, it was predicted that a computer program capable of beating a human expert Go player was at least a decade away.

As DeepMind researchers elaborated in a January article in Nature, AlphaGo's expertise is the result of a combination of Monte-Carlo tree search (an algorithm for optimal decision making) and deep neural networks that have been trained via supervised learning, observing human expert games, and reinforced by playing games against itself.

"To be honest we are a bit stunned and speechless," said DeepMind founder Demis Hassabis in a post match press conference. "AlphaGo can compute tens of thousand positions a second, but it's amazing that Lee Sedol is able to compete with that and push AlphaGo to the limit. We came here to challenge Lee Sedol because we wanted to see what AlphaGo was capable of, and his amazing genius and creative skills have done that."

While a significant milestone for AI research, AlphaGo's Saturday win was unfortunate for Sedol as he is no longer eligible for the $1 million series prize. Nevertheless, Sedol will play the remaining two matches on Sunday and Tuesday afternoon in Seoul for the chance to win £15,000 for each victory.