After winning the fifth and final game in the Deep Mind Challenge, AlphaGo emerged victorious against Lee Se-dol, a 9-dan human Go player. AlphaGo now has an honorary 9-dan ranking.

Go, a game in which there are more moves than there are atoms in the universe, had until recently been considered the final frontier for artificial intelligence. Whereas Chess only needs brute force computation Go was regarded as a game that needed the sort of thinking that only humans possess. The fact that Google's AlphaGo, a neural network that combines reinforcement learning with backpropation, has now beaten one of the world's best Go players is a breakthrough.

It was after AlphaGo beat the European Go Champion, see Google's AI Beats Human Professional Player At Go, that Google arranged for a tougher test of the software in which AlphaGo would take on the 18-times world champion, South Korea's Lee Se-dol. The five game match with a prize of $1 million took place in Seoul, starting on March 9, 2016 and by March 12 AlphaGo had won 3-0 by taking the first three games of the series. According to the rules of the match, however, all five had had to be played. Lee Se-dol won the fourth game in which he made a brilliant move that caused AlphaGo to make a mistake and commented afterwards on the fact that had never been congratulated so much for simply winning a single game! See AlphaGo Has Lost A Game - Score Stands At 3-1 for an account of the match up until that point.

You might imagine that with the match already won, there would be less interest in the final game on March 15. Far from it - South Korea's normal public broadcasting service was suspended in order to livestream the game.

A livestream was, and still is, also available on the Deep Mind You Tube Channel. Meanwhile Demis Hassabis, CEO of Deep Mind and therefore the person widely credited with AlphaGo's success, made several tweets to over 25K followers. Prior to the start of the match:

The game turned out to be a cliffhanger as indicated by this tweet, made part way through, he notified of an unexpected error made by the machine:

Then two hours later came two more tweets:

In recognition of its achievement of beating a 9-dan player, AlphaGo has been awarded the same ranking which is the highest Go grandmaster rank and, according to South Korea's Go Association is reserved for those whose ability at the ancient board game borders on "divinity",

The scoreboard show the final position and the fact that after 280 moves AlphaGo achieved a win by resignation.  

 The Goole Asia Pacific Blog report provides some more details:

After a loss in Game 4, and a move early on that looked like a mistake, but could have been a creative and effective new move, AlphaGo won Game 5 against the legendary Lee Sedol. This final game of the match was close until the very end, with commentators going back-and-forth about who was on top. But after 280 moves, Lee resigned, having gone through two byō-yomi overtime periods.

It quotes Michael Redmond the American commentator who covered the entire match as saying:

“It was difficult to say at what point AlphaGo was ahead or behind, a close game throughout. AlphaGo made what looked like a mistake with move 48, similar to the mistake in Game Four in the middle of the board. After that AlphaGo played very well in the middle of the board, and the game developed into a long, very difficult end game.

Commenting on the match as a whole Kim Seongryong, 9-dan, a Korean commentator said: 

“Just like the scientists, Go players are always trying to find new methods and approaches. And we are so happy when we find them. This Challenge Match has brought us Go players to new areas we’ve never explored. We are now seeing a lot more interest in playing Go. And even in one week, I feel like my Go playing has improved.”

It is not just Go that is affected by this historic breakthrough. As explained by Mike James in Why AlphaGo Changes Everything, given that AlphaGo which can turn its abilities to other areas this achievement signals a new era for AI as a whole. 

More Information

AlphaGo’s ultimate challenge: a five-game match against the legendary Lee Sedol

Mastering the game of Go with deep neural networks and tree search

Related Articles

Why AlphaGo Changes Everything

AlphaGo Has Lost A Game - Score Stands At 3-1

Google's AI Beats Human Professional Player At Go

To be informed about new articles on I Programmer, sign up for our weekly newsletter, subscribe to the RSS feed and follow us on, Twitter, Facebook, Google+ or Linkedin. 

Comments

Make a Comment or View Existing Comments Using Disqus

or email your comment to: comments@i-programmer.info