News / Articles

Game Over? AlphaGo Beats Pro 5-0 in Major AI Advance

[sgf sgfUrl="http://www.usgo.org/news/wp-content/uploads/2016/01/AlphaGo1.sgf" href="http://www.usgo.org/news/wp-content/uploads/2016/01/AlphaGo1.sgf" class="alignright"] [/sgf]
In a stunning development, the AlphaGo computer program has swept European Go Champion and Chinese professional Fan Hui 2P 5-0, the first time that a go professional has lost such a match. "This signifies a major step forward in one of the great challenges in the development of artificial intelligence - that of game-playing," said the British Go Association, which released the news on January 27, based on findings reported in the scientific journal Nature this week (click here for the video, here for Nature's editorial, Digital intuition and here for Go players react to computer defeat). NOTE: This story was posted at 1p EST on Wednesday, January 27; be sure to get the latest breaking go news by following us on Facebook and Twitter. "AlphaGo's strength is truly impressive!" said Hajin Lee, Secretary General of the International Go Federation and a Korean go professional herself. "Go has always been thought of as the ultimate challenge to game-playing artificial intelligence," added Thomas Hsiang, Secretary General of the International Mind Sport Association and Vice President of International Go Federation. "This is exciting news, but bittersweet at the same time," said American Go Association president Andy Okun. "I think we go players have taken some pride in the fact that we could beat the best computers. Now we're down to Lee Sedol fighting for us." Google DeepMind, the British artificial intelligence company which developed AlphaGo, has issued a challenge to Lee Sedol 9P from South Korea, the top player in the world for much of the last 10 years, to play a 5-game, million-dollar in March. "I have played through the five games between AlphaGo and Fan Hui," said Hsiang. "AlphaGo was clearly the stronger player. The next challenge against Lee Sedol will be much harder." While Hajin Lee agreed, saying "I still doubt that it's strong enough to play the world's top pros," she added "but maybe it becomes stronger when it faces a stronger opponent." Fan Hui (left) is a naturalized French 2-dan professional go player originally from China. European Champion in 2014 and 2015, Fan is also a 6-time winner in Paris as well as Amsterdam. Just as the Kasparov/Deep Blue match did not signal the end of chess between humans, "so the development of AlphaGo does not signal the end of playing go between humans," the BGA pointed out. "Computers have changed the way that players study and play chess (see this 2012 Wired article), and we expect something similar to occur in the field of go, but not necessarily as assistance during play. It has been recognised for a long time that achievements in game-playing have contributed to developments in other areas, with the game of go being the pinnacle of perfect knowledge games."  Added Okun, "go has for thousands of years been a contest between humans and a struggle of humans against their own limits, and it will remain so. We still cycle in the Tour de France, even though we've invented the motorcycle." The BGA noted that that achievements in game-playing technology have contributed to developments in other areas. The previous major breakthrough in computer go, the introduction of Monte-Carlo tree search, led to corresponding advances in many other areas. Last year, the Facebook AI Research team also started creating an AI that can learn to play go and earlier today Mark Zuckerberg reported on Facebook that “We're getting close, and in the past six months we've built an AI that can make moves in as fast as 0.1 seconds and still be as good as previous systems that took years to build. Our AI combines a search-based approach that models every possible move as the game progresses along with a pattern matching system built by our computer vision team.” In a related story, computer scientist John Tromp last week revealed the number of legal go positions, "weighing in at 9*19=171 digits." Read more here. Game 1 of the AlphaGo vs. Fan Hui 2P match appears above right. Click below for the match's remaining game records: AlphaGo vs. Fan Hui, game 2 AlphaGo vs. Fan Hui, game 3 AlphaGo vs. Fan Hui, game 4 AlphaGo vs. Fan Hui, game 5
Update (11:44pm 1/27):Myungwan Kim 9P will analyze the games played between Fan Hui and AlphaGo during a live stream on the AGA YouTube Channel and TwitchTV this Friday; more details will be posted at 7a EST.