Google DeepMind’s AlphaGo scored another win Thursday against star South Korean player Lee Se-dol, giving a boost to Google’s growing position in the field of artificial-intelligence.
The close contest in Seoul was the second in a series of five games being played in South Korea, the last of which is scheduled for Tuesday.
On Wednesday, a shocked Lee had resigned the first game to AlphaGo in a cliffhanger.
AlphaGo appears to have surpassed its first day's play with Lee saying that if he was surprised on Wednesday, he was now "speechless." From the beginning of the game, there was not a moment when he felt he was leading in the game, Lee told reporters. In the game Thursday, AlphaGo's moves were flawless unlike in the previous day when some moves were problematic, he said.
The program now needs just one more win to claim victory in the tournament. The third game won't be easy, but Lee will try hard by focusing on the beginning part of the game, he said.
Analysts have been curious about how AlphaGo tackles the board game, which is said to require a lot of intuition and feel rather than just brute calculation, as players take turns to place black or white pieces, called “stones,” on the 19-by-19 line grid to capture the opponent's stones by surrounding them and encircling more empty space as territory.
In a match in October, when AlphaGo won 5-0 against an European Go champion Fan Hui, the program made brilliant moves but several critical mistakes, which made it look very much like an artificial program or computer, said Michael Redmond, commentator at the Seoul event and a professional Go player. Redmond said he would be watching the games in this match closely to see if AlphaGo would again make critical mistakes.
Google DeepMind has said that unlike humans the AlphaGo program aims to maximize the probability of winning rather than optimizing margins, which helps explain some of its moves.
The contest has been watched online by a large number of users, who are fans of Go or are intrigued by the epic man vs. machine battle playing out in Seoul. GoogleMInd said 95,000 viewers watched the first game live with another 1,250,259 viewers watching the recording.
For Google a win by AlphaGo in the complex Go game would parallel if not surpass the chess victory of IBM's Deep Blue against Garry Kasporov in 1997 and the 2011 win in the Jeopardy quiz show by Watson, another computer from Big Blue.
Watson went on to become a technology that IBM has commercialized to offer natural language processing and machine learning for the analysis of unstructured data, and Google’s bid to position itself as a key winner in AI could also be directed at boosting its presence in this market.
“So far, IBM has really stolen the show when it comes to mass market awareness of cognitive computing via AI,” said Patrick Moorhead, president and principal analyst of Moor Insights & Strategy.
“Both Google and Microsoft have advanced AI capabilities, so yes, Google's AlphaGo victory is a way to raise awareness of the capability, portending to an increased AI offering to businesses,” Moorhead added.
There have been reports, however, that Lee is not the current strongest Go player, a position that is attributed to Chinese player Ke Jie. Lee has 18 world titles since he went professional in 1995, but Ke is undefeated against Lee, having beaten him twice this year, according to The Wall Street Journal. In 2015, he was the first player after Lee to win two world titles in one year, it added. Ke has won 173 of 247 games he played, giving him a winning percentage of 70 percent, while Lee has won 68.2 percent of 1173 games played, according to website Go4Go.