Google-powered AI finally beats human champion in ancient Chinese game Go
The intelligent machine created by Google’s DeepMind team first competed with other programs and managed to win all but one of its 500 games, Google writes in its blog.
What is Go? It’s a 2,500 year-old board game that’s nearly impossible to beat using artificial intelligence. pic.twitter.com/UEyGIxh42I— Google (@google) January 27, 2016
The next step was to invite an actual human to challenge the AI. This is where reigning three-time European Go champion Fan Hui stepped in. Google describes him as “an elite professional player who has devoted his life to Go since the age of 12.”
The intellectual battle between the human and the system took place in October and AI didn’t leave any chance to Fan Hui. The report on the epic Go battle was released only on Wednesday in the journal Nature.
“In a closed-doors match last October, AlphaGo won by 5 games to 0. It was the first time a computer program has ever beaten a professional Go player,” Google said.
The game of Go, which originated in China more than 2,500 years ago, was not just picked by chance to test the AI.
According to Google, Go is a game of “profound complexity” and played “primarily through intuition and feel.” Also there are billions and billions of possible positions—“that’s more than the number of atoms in the universe.”
“This complexity is what makes Go hard for computers to play, and therefore an irresistible challenge to artificial intelligence (AI) researchers, who use games as a testing ground to invent smart, flexible algorithms that can tackle problems, sometimes in ways similar to humans.”
However, AlphaGo is not going to stop. Next time it will face one more Go guru – the legendary Lee Sedol, “the top Go player in the world over the past decade.”
Google says that the company is “thrilled” to have mastered legendary Go and “thus achieved one of the grand challenges of AI.”
“Because the methods we’ve used [in AlphaGo] are general-purpose, our hope is that one day they could be extended to help us address some of society’s toughest and most pressing problems, from climate modeling to complex disease analysis.”
The first game ever mastered by a computer was noughts and crosses (tic-tac-toe) back in 1952. Some forty-five years later, in 1997, computer Deep Blue beat famous Russian Grandmaster Garry Kasparov at chess.