Google AI AlphaGo Defeats South Korean Go Master

Updated on

AlphaGo, a computer program developed by Google has beaten a South Korean master of the ancient board game Go.

The duel was billed as the match of the century, testing the Google artificial intelligence against one of the best human players of the famously complex board game. The match was concluded Tuesday, with the program AlphaGo running out 4-1 victor against Lee Se-dol.


Google AI AlphaGo beats human Go champion

Computer programmers are trying to make software that can beat humans, and this victory is a major step.

“It made me question human creativity. When I saw AlphaGo’s moves, I wondered whether the Go moves I have known were the right ones,” the human competitor, Lee Se-dol, 33, said during a postmatch news conference. “Its style was different, and it was such an unusual experience that it took time for me to adjust.”

“AlphaGo made me realize that I must study Go more,” said Mr. Lee, one of the world’s most accomplished players.

Hugely complex game invented over 3,000 years ago

The board game involves two players and is thought to have been invented in China over 3,000 years ago. The strategy game involves competing for territory by putting black and white stones on a 19 x 19 board.

Go has proven to be a tough hurdle for artificial intelligence to beat humans. IBM supercomputer Deep Blue beat world champion Garry Kasparov at chess in 1997. Despite predictions that it would take the program another 10 years to be able to beat a Go champion like Lee, the artificial intelligence smashed expectations.

AlphaGo was developed by DeepMind, Google’s artificial intelligence company. The program beat three-time European Go champion Fan Hui 5-0 in October before taking on the stronger Lee, who has won 18 international titles.

AlphaGo won three games in a best-of-five. Lee won the fourth game on Sunday but could not prevent AlphaGo winning a $1 million prize which Google is set to donate to Unicef and other organizations.

Lee finds his passion for the game reignited after matches

The game is most popular in Northeast Asia, and many fans tuned in to watch the match. The tension was evident, with both players using more of the allotted one minute to think about moves before placing a stone.

Hong Seok-hyun, head of the Korean national Go association, awarded AlphaGo the honorary certificate of the Go degree of Nine Dan, the highest level. Lee also holds the degree of Nine Dan.

Demis Hassabis, the chief executive of Google DeepMind, said Lee had made the team aware of some weaknesses in AlphaGo. Hassabis said that the algorithms used in AlphaGo “one day can be used in all sorts of problems, from health care to science.”

Over 100 million people watched the software battle it out against Lee. Hassabis hopes that the widespread attention will drive more people to learn Go, which he calls the “most profound game humankind has devised.”

Lee claimed that AlphaGo artificial intelligence used strategies unlike any he had ever seen in human opponents. “It remained unfazed psychologically and stayed focused,” he said. “In that regard, I don’t think humans can beat it, even though I hesitate to admit that AlphaGo is above humans in Go skills yet.”

Lee was apparently questioning his love for Go prior to the matches with AlphaGo, but he said that the recent matches had renewed his passion for the game. He has been playing professionally since age 12.

“I have some regrets about the matches I have played against AlphaGo,” he said. “But I could not have enjoyed them more.”

Leave a Comment