Roundup: AlphaGo defeats Lee Sedol 4-1 in historic Go match

Google's Go-playing computer program AlphaGo on Tuesday ended a historic match of the ancient Chinese board game with Go grandmaster Lee Sedol of South Korea by taking a 4-1 lead with its fourth victory in the final match of the best-of-five series.

The final winner was already determined before Tuesday's encounter as Lee lost the first three games of the five-game match. AlphaGo got 1 million U.S. dollars in prize, which will be donated to charities.

The human Go champion beat the artificial intelligence (AI), developed by Google's London-based AI subsidiary DeepMind, in the fourth match, but Lee was defeated once again in the final match.

Despite the sweeping victory, it is too early to say that AI has surpassed humans in Go because Lee fought AlphaGo with little knowledge about the computer program, including a playing style and a strategy, one commentator said.

Only after analyzing AlphaGo's strategy sufficiently and holding a match once again between the AI and human Go players, can it be determined whether machines may surpass humans in the board game, the commentator said.

Lee and AlphaGo exchanged 280 moves for five hours, the longest among the five games that kicked off last Wednesday in Seoul. Lee displayed a fighting spirit to the last minute, but he eventually lost by a narrow margin.

The 33-year-old kept a tight game with the two-year-old computer program by the middle of the match, but a decisive moment came past halfway as AlphaGo succeeded in attacking the lower-left side.

At the lower-left side, AlphaGo made questionable moves, which looked like blunders at the first sight, but those also proved to be a strategy to win more territory in hindsight.

Commentators said AlphaGo made excellent moves in the latter half of the game, proving its better capability to play as time goes on. Lee didn't make any big missteps during the game, experts said.

About four and a half hours into the match, AlphaGo had been subject to the one-minute countdown for the first time as the Go-playing AI consumed all of the given two-hour time limit.

AlphaGo made a very rapid decision on moves, in which human Go players tend to take very long, but the computer program took a long time to make moves that humans think of as easy to determine, according to commentators.

Playing black, Lee put his first two stones right beside flower spots in the right side, while AlphaGo placed its first two white stones in flowers in the left side.

After his first win on Sunday, Lee offered to play black stones, with which he believed AlphaGo displayed a relatively weak play.

About two and a half hours into the match, a tight game continued. AlphaGo attacked the right center by seeking to build a large territory there, while Lee built a territory in the lower-right side of the board according to his strategy which reportedly occupies more areas in an early phase.

In the fourth match where Lee won his first victory after three straight losses, AlphaGo made bad moves after Lee began to dominate in building a territory.

Lee's first victory over AlphaGo indicated that AI hasn't surpassed humans completely in Go, which had been regarded as the last game humans can dominate over machines due to its complex, intuitive and creative nature.

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 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.

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 territory on the grid. One can remove stones of the opponent by surrounding the pieces. Endit