On Tuesday, DeepMind’s AlphaGo defeated the current top-ranked player, Ke Jie, by 0.5 points. The match was the first of three encounters between Google’s AI and the 19-year-old, and they will keep facing each other at The Future of Go Summit in Wuzhen, China.
The week-long event will also feature press conferences and forums on artificial intelligence.
280 million people witnessed how AlphaGo defeated Lee Sedol, the best Go player at the moment, back in March last year. The matches framed in The Future of Go Summit are also being live streamed on YouTube, with an average duration of six hours each.
Ke Jie copied moves used by AlphaGo before
Tuesday’s match was not the first time Ke Jie faced AlphaGo. The two master players played against each other online earlier this year when DeepMind tested a new version of the AI against world-class strategists.
Go is a millenary strategy board game, in which typically two players battle to control space on the board. The pieces are called stones and players lay them in the intersections of a 19 x 19 grid.
During that early encounter, Ke Jie witnessed how AlphaGo made some unorthodox moves to win, which since then have been increasingly adopted by more and more players as they understand the thinking behind the brain powering the AI.
This time around, however, the 19-year-old player tried to take the lead and literally beat AlphaGo at its own game using some of its signature moves. He started early on with two 3-3 moves to secure some corners of the board, but Google’s AI managed to turn the tables and end up winning by a 0.5 difference.
AlphaGo is changing the millenary game
Ke Jie and AlphaGo are set to play a three-match series, with the next two games happening on Thursday and Saturday respectively. In between, the AI won’t rest, as it engages in some interesting modalities of Go with other human masters.
Gu Li and Lian Xiao, two world-class Go players that challenged AlphaGo back in January, will each team up with Google’s AI as their partner in a couple of matches.
Then, on Friday, AlphaGo will take on Chen Yaoye, Zhou Ruiyang, Mi Yuting, Shi Yue, and Tang Weixing all at once in a five vs. one. These alternative game modes attempt to prove how quantitatively capable is DeepMind against such a concerted brain effort.
If you don’t feel like watching the entire six hours of the event, DeepMind has an interactive Go board in which you can select the matches and replay them move by move as they happened in real life.
Source: DeepMind