Go to Contents Go to Navigation

(LEAD) Google AI marks third win against S. Korean Go champion

All News 18:15 March 12, 2016

(ATTN: ADDS details, comments, photo)

SEOUL, March 12 (Yonhap) -- Google's artificial intelligence (AI) program AlphaGo celebrated its third consecutive victory against South Korean Go player Lee Se-dol on Saturday, heralding another win for machines against mankind.

In their third showdown at the Four Seasons Hotel in Seoul, AlphaGo, the AI designed by Google's London-based firm DeepMind, collected another win, beating Lee by resignation in 176 moves. The self-learning program had already won two games of the five-round tournament earlier this week.

By taking the first three matches, AlphaGo is assured of a series win and US$1 million in prize money, which will be donated to UNICEF and other charities.

Lee is a ninth-dan player who went pro at the age of 12 and has won 18 international events. European champion Fan Hui, a second-dan player, lost five consecutive matches against AlphaGo last October.

Lee will get two more cracks at beating the software on Sunday and on Tuesday, at the same venue.

Taking a lesson from previous games, Lee started Saturday's showdown with an aggressive strategy, hoping to snatch early control. Lee Hyeon-wook, an eighth-dan player providing commentary on the match, also said Se-dol appeared to have gone back to his "audacious, feisty" teenage years, after losing the first two games playing a more conservative style.

AlphaGo met the challenge with its sleek defense and ultimately forced Lee into resignation.

South Korean Go players watching the game noted that AlphaGo appeared to be playing the percentages to great effect.

"AlphaGo went by the book early on and then showed some unconventional, unpredictable moves later," said Seo Keon-woo, a sixth-dan player who will referee the fourth game. "When AlphaGo finds itself in an advantageous position, it tries to protect its territories."

Lee Hyeon-wook said AlphaGo doesn't tend to take risks.

"AlphaGo doesn't necessarily try to win by a large margin; it does just enough to win, period," Lee added. "It appears to be programmed to minimize risks."

Kwon Kap-yong, an eighth-dan player and a former teacher of Lee Se-dol's, said AlphaGo played "a consistent game" from start to finish, while his former pupil showed some vulnerability.

"Lee Se-dol is human after all," Kwon said. "AlphaGo had the edge as the game wore on."

Go, known as "baduk" in Korea, originated in China more than 2,500 years ago. It involves two players alternately putting black and white stones on a checkerboard-like grid of 19 lines by 19 lines. The object is to claim larger territories than one's opponent by surrounding vacant areas of the board using one's own stones.

Go has been viewed as one of the hardest games for computers to master. Google said the possible number of board configurations of Go is larger than the number of atoms in the universe.


Send Feedback
How can we improve?
Thanks for your feedback!