AI wins as Google algorithm beats No. 1 Go player

As we reported last month, the Future of Go Summit is a collaboration between Google, the China Go Association, and the Chinese government created to explore aspects of Go as played by humans and AI and to test the limits of both. The first, Pair Go, will see two Chinese pros face off against each other alongside an AlphaGo teammate: since Go strategy requires vast foresight and strategy, the helping hand could be more of a hindrance in practice for both human and computer, removing their abilities to plan ahead.

AlphaGo is taking on 19-year-old world champion Ke Jie in a three-game match that spans five days this week, and it won the first game handily on Tuesday.

After Tuesday's match, Ke will play against AlphaGo on Thursday and Saturday, each game lasting seven hours. While the grander drama plays out, a $1.5 million prize is also at stake.

AlphaGo's machine-learning algorithm integrated advantages of both a "policy network" and a "value network", storing not only innumerable past games played by humans but also those played against the continuously improved versions of itself. In the pair Go match, Chinese Go player Gu Li will compete with the other player Lian Xiao on May 26, with each pairing up with his AlphaGo teammate. Will the collaborative approach help the human players by providing an avenue for new ideas, or will it be a case of too many people cluttering the strategy?

The program defeated South Korean Go master Lee Se-dol 4-1 in March 2016.

Also like a god, AlphaGo has been demonstrating its influence far and wide among the game's top players - including Ke, who Wired notes adopted some tactics from AlphaGo.

In the first of three planned games in the eastern water town of Wuzhen, the AlphaGo program held off China's world number one Ke Jie in front of Chinese officials and Google parent Alphabet's chief executive Eric Schmidt.

Go, an ancient Chinese board game, is favored by AI researchers because of the large number of outcomes compared to other games such as western chess. But rather than cowering in fear or kneeling before the dominance of AlphaGo, Deep Blue, Watson, and all of the other computers intent on literally beating us at our own games, it's time for us to start fighting back. "I will never give up", Ke said.

"Last year, it was still quite humanlike when it played", Mr. Ke told the New York Times after AlphaGo's win on Tuesday.

"If you had been asking me a few years ago, how long before we see a computer program strong enough to beat the best human at Go, I was thinking it would be not in my lifetime", said Ryan Hayward, a professor in the U of A's computing science program. Aside from the foray in Go, the AlphaGo program is entering healthcare next, playing an active role in the research and treatment of complicated diseases including diabetes and cancer, said Shi Bomeng, President of Google China. The game can't be found on numerous most-watched streaming services in China, such as Tencent Video and Sina.com.

  • Joe Gonzales