AlphaGo clinches victory for machine over man in tough Chinese game

Google's program has defeated the world Go champion in three of the five games.

In a decisive and consecutive third win over professional Go player Lee Sedol, Google's computer program AlphaGo has shown that man has been outwitted by his own creation.

The 'digital intuition' displayed by the machine in winning the complex ancient Chinese game proves that artificial intelligence systems are capable today of decision making and winning in ways that may prove tough to be predicted.

Lee lost the third game very early and AlphaGo won the game convincingly enough for its critics and doubters. At the press conference, Lee Sedol confessed to feeling pressure like never before and while he had identified missed opportunities in game 2, he felt he could not win game 1 even if he replayed it today.

It now remains to be seen if he will manage to win even a single game in the five-game challenge. It is being felt that his only chance would be to pull the plug on AlphaGo, writes Nature.

Lee had studied the two first games with a group of fellow Korean professional players and practised every tactic to stay calm and relaxed. But clearly, it did not help. Even the sharp strategic battles he engaged the computer in came to nought.

Very gradually onlookers realized how Lee went from attack to defence and AlphaGo's responses made it feel like "it knew exactly what Lee's plan was". Strong and detached, the computer worked on winning rather than maximising advantage. Tired and defeated after 176 moves, Lee conceded defeat.

Where the computer scores is also in the fact that its knowledge of the game and its moves is implicit and emerges from its statistical comparisons of types of winning board positions at Go. This makes it difficult to predict or know the mind of the computer, adding to fears of the machines taking control of humans.

AlphaGo's triumph "shows that the methods we do have are even more powerful than we first thought," said AI expert Stuart Russell of the University of California's Berkeley Electrical Engineering & Computer Sciences department. "The fact that AI methods are progressing much faster than expected makes the question of the long-term outcome more urgent," he told AFP.

Related topics : Artificial intelligence