Google's AlphaGo takes lead over human champion

March 9, 2016 | 15:37

Tags: #ai #alphago #artificial-intelligence #fan-hui #go #lee-se-dol

Companies: #deepmind #google #research

Google's DeepMind-powered AlphaGo artificial intelligence has successfully defeated champion Lee Se-dol in the first of a scheduled five matches, marking the first time a computer has beaten a player of his level.

Google unveiled AlphaGo back in late January as a demonstration of the possibilities promised by its DeepMind platform. Go has been a mainstay of artificial intelligence research for years, thanks to its complexity: where chess, which has had computers beating grand masters for some time, has 10-to-the-power-of-120 games, Go has 10-to-the-power-of-761. The result: brute-force solutions are out of the question, leaving researchers having to find ways to imbue programs with intelligence.

To prove the program's prowess, Google asked European Go champion Fan Hui to test it out. The results astounded AI researchers the world over: AlphaGo won five matches to zero, a world first. Now the company is facing a tougher challenger: 9-dan champion Lee Se-dol in a series of five matches, and things aren't looking great for the human side of the table. In the first match, AlphaGo once again triumphed with Lee conceding to his electronic opponent with just under half an hour left on the game clock.

'I was very surprised,' The Verge reports Lee saying after the match. 'I didn't expect to lose. I didn't think AlphaGo would play the game in such a perfect manner. I enjoyed this game and look forward to the next. I think I failed on the opening layout so if I do a better job on the opening aspect I think I will be able to increase my probability of winning.'

Lee will play another four matches, scheduled for late tonight, Saturday, Sunday, and Tuesday, against AlphaGo, with a $1 million prize at stake for the overall winner.
Discuss this in the forums
YouTube logo
MSI MPG Velox 100R Chassis Review

October 14 2021 | 15:04