Google DeepMind’s AlphaGo AI has a pretty good track record at the ancient Chinese game of Go. Now, the AI has defeated Ke Jie, who’s widely-regarded as the world’s best (human) Go player.
Although he lost in the end by just half a point (AlphaGo tends to win by small margins, instead focusing on making moves that will lead to higher chances of a win rather than higher margins of victory), Ke Jie played pretty cleverly against the AI — at least aware of the tactics that DeepMind’s creation has been using to conquer dozens of other players over the last couple of years.
“Interesting that Ke Jie has decided to play an very early 3-3 point as he knows
#AlphaGo likes to play there,” DeepMind founder Demis Hassabis tweeted near the beginning of the game. “Ke Jie is using the ideas #AlphaGo used in the master series of online games in January against AlphaGo :) intriguing to see what it will do,” he said in another tweet.
Today’s game was one of three happening between AlphaGo and Ke Jie this week at the Google-hosted Future of Go Summit in Wuzhen, China. Besides the three main games, the schedule for the event also mentions a big team game where two Go masters will face off both using AlphaGo as a teammate, and another where AlphaGo will take on five humans at once.
The other two major games between Ke Jie and AlphaGo are to follow this week, with one on Thursday and one on Saturday (Wednesday and Friday, respectively, in the US). If you want to watch a replay of all of the moves from the first match between Jie and AlphaGo (and later, the other games), head over to the Future of Go Summit’s website.
On a related note, China reportedly censored the livestream of the game:
Still, China showed some skittishness at game time. Despite huge interest, many Chinese became consternated when it became apparent there was no obvious live video of the event online. A site that follows Chinese censorship orders, China Digital Times, posted a translated notice from the government calling for all websites to block the broadcast.
FTC: We use income earning auto affiliate links. More.