Human Go champion Lee Sedol of South Korea finally defeated Google's Go-playing artificial intelligence (AI) AlphaGo on Sunday in the fourth match after three straight losses at the ancient Chinese board game. (Xinhua photo)
Human Go champion Lee Sedol of South Korea finally defeated Google's Go-playing artificial intelligence (AI) AlphaGo on Sunday in the fourth match after three straight losses at the ancient Chinese board game.
The final winner of the best-of-five match was already determined as the Go-playing program won the third consecutive game on Saturday, but Lee raised his score to 3-1 with the first victory over the supercomputer.
The "match of the century" drew attention from around the world as it is seen as a representative match between humans and AI.
Lee's victory defended the last remaining pride of human Go players as he achieved his goal of winning at least one game.
AlphaGo, a computer program developed by Google's London-based AI subsidiary DeepMind, took the world by surprise as it beat European Go champion Fan Hui by 5-0 in October last year and defeated Lee at three straight games.
AlphaGo boasts of a deep learning capability to learn for itself and discover new strategies by playing games against itself and adjusting neural networks based on a trial-and-error process known as reinforcement learning.
Lee, 33, is regarded as one of the greatest Go players in the world as he won 18 world championships for 21 years of his professional career.
He recorded a winning rate of about 70 percent with 47 victories in professional matches.
One commentator said Lee's victory was a proof that AI hasn't surpassed humans completely, and another commentator said the fourth match revealed AlphaGo's weak points also shown in the previous game.
The first victory of the human champion came as AlphaGo made many "questionable" moves in the middle of the match, which commentators said must be mistakes and algorithm errors.
Such moves, which were also displayed in the previous matches, proved to be a strategy for a "bigger picture" in hindsight that had not been found in the human Go world, so it made the match unpredictable until near the end.
Playing white, Lee adopted a strategy in an early phase to occupy the left and right sides of the board, while AlphaGo sought to win a broader area at the center.
The 33-year-old attacked the center and made excellent moves, which drove AlphaGo into pressure and made many questionable moves.
Keeping his dominating position, Lee eventually defeated AlphaGo as the AI resigned after exchanging 180 moves for nearly five hours in the fourth match that started from 1:00 p.m. local time at Four Seasons hotel in Seoul.
The final match will be held in Seoul on Tuesday after a one-day break.
Go, known as weiqi in China and Baduk in South Korea, originated from China thousands of years ago.
It involves two players who take turns putting white and black stones on a grid of 19 lines by 19 lines.
One can win an opponent when gaining more intersections on the grid. One can occupy stones of the opponent by surrounding the pieces.
