Subscribe
E-Paper
 |
| Lee Se-dol smiles while speaking during an interview after winning a Go game against HanDol, a locally developed artificial intelligence program, in Seoul, Wednesday. /Yonhap |
By Kim Se-jeong
Korea's all-time Go master Lee Se-dol beat won a game against an opponent powered by a locally developed artificial intelligence program in Seoul, Wednesday, three years after his historic match with Google's AlphaGo in 2016.
Lee won the first round of the match against HanDol, a program developed by NHN Entertainment Corp. Two more games are scheduled for today and Saturday.
Go, known as baduk in Korean, is a strategy game originated in China 3,000 years ago.
In the 2016 match, Lee beat Google AlphaGo's DeepMind in one out of five matches. In November, the 36-year-old announced his resignation, saying the life of a Go master would be meaningless unless he can beat the AI.
"I am not at the top even if I become the No. 1. There is an entity that cannot be defeated," Lee told Yonhap News Agency during an interview last month. At the invitation of HanDol, Lee delayed his official resignation.
HanDol has already beaten five players in Korea before Wednesday's match.
"I didn't imagine the game to be easy. I should be happy but I feel a little bit empty," Lee said. "HanDol has a homework to do for the next two rounds."
Commentators of the game said HanDol made some unusual mistakes during the game. They also said Lee's unusually defensive strategy worked and the 78th move was critical.
In the nationally televised game, Lee looked puzzled at times and frustrated by the AI's moves.
The game had the same format as seen in the 2016 match.
An amateur 5-dan player Lee Hwa-seop placed the stone on the physical board on behalf of HanDol. "Lee Se-dol's match is significant on so many levels. I feel pressure on my shoulders, but I will try to enjoy it," he said to Yonhap.
Lee began playing Go at the age of five and became a professional player seven years later. He won championships 18 times.
Also popular in Japan and Korea, the game has developed slightly different rules in each country. The strategy game has its players across the world, and the three countries have long been in competition to dominate the game's global standing.
Lee's defeat in 2016 was significant for artificial intelligence.
"On behalf of the whole AlphaGo team at DeepMind, I'd like to congratulate Lee Se-dol for his legendary decade at the top of the game, and wish him the very best for the future," Demis Hassabis, co-founder of DeepMind, said then.
