Can a machine beat a human? That’s what Lee Se-dol, the world champion of the ancient Chinese board game Go, is about to find out.
Lee is facing AlphaGo, a computer program built by Google DeepMind, on Wednesday in the first of five games of Go.
The stakes are high. If Lee wins, he gets $1 million and reasserts his title as global champ; a loss by the South Korean would signal another technological advance by scary smart computers. (Google will donate the prize money to charity.)
“I heard Google DeepMind’s [artificial intelligence] is surprisingly strong and getting stronger, but I am confident that I can win,” Lee said before the match in Seoul.
Lee, 33, holds the highest possible professional ranking for a Go player — 9 dan — and has been called “the Roger Federer of Go.”
Go originated thousands of years ago in China. Two players take turns placing black and white stones on a square grid of 19 lines by 19 lines; the goal is to take territorial control of the board by using pieces to surround the opponent. Games can last for hours, and winning requires immense mental stamina, intuition and strategy.
Teaching computers to master Go has been kind of a holy grail for artificial intelligence scientists. There are more possible configurations of the board than there are atoms in the universe, said Demis Hassabis, CEO of Google DeepMind, which developed AlphaGo.
“Go is the most profound game that mankind has ever devised,” Hassabis said. “Go is a game primarily about intuition and feel rather than brute calculation, which is what makes it so hard for computers to play well.”
Last October, AlphaGo convincingly defeated the European Go champion, Fan Hui, obliterating him in five consecutive matches. The computer’s victory was considered a huge breakthrough, occurring roughly a decade sooner than experts had expected.
Back in the 1990s, software programs became adept at classic board games like backgammon. Their rapid progress culminated in the historic victory of IBM’s Deep Blue computer over world chess champion Gary Kasparov in 1997.
But it’s taken another two decades for artificial intelligence to get to grips with the mind-boggling complexities of Go. Until recently, software programs could only compete with human amateurs.
Google researchers say they expect AlphaGo’s attributes to eventually be put to use beyond the realm of board games in areas like Google’s own apps, and even medicine.
Lee’s matches against AlphaGo run through March 15, and are being live streamed on DeepMind’s YouTube channel, and broadcast on television throughout Asia.
Google acquired DeepMind in 2014 in efforts to beef up its portfolio in artificial intelligence and robotics.