AlphaGo beats Lee Sedol in first Google DeepMind Challenge match

The first of the five-match DeepMind Challenge series between Google's AI AlphaGo and world champion Lee Sedol saw the algorithm beat him in the Go board game.

 

 

In a victory that will go down in the history books, Google’s Go-playing DeepMind algorithm AlphaGo defeated legendary South Korean Go-player, Lee Se-Dol. The tournament, as an exhibition of the progress of artificial intelligence, is being held in Seoul, South Korea, and will see Lee Se-Dol, a professional 9-Dan Go player, take on a program in the board game. The game lasted for about three and half hours, after which Lee resigned to give AlphaGo a historic victory and set a landmark for achievements in the field of artificial intelligence. The tournament has Lee participating for a $1 million prize, and after the first match, he expressed his surprise at the level of perfection and skill at which AlphaGo played.

AlphaGo, an algorithm developed by Google’s R&D laboratories, uses a system of deep neural networks and machine learning techniques to assess and make moves. Go, an ancient and severely complex Chinese board game, was seen as a tasking premise to test the algorithm’s development, and at the historic opening game held today, AlphaGo excelled with a surprisingly strong opening formation along with a few unexpected moves to defeat Lee. “I was very surprised," said Lee, "I didn't expect to lose, I didn't think AlphaGo would play the game in such a perfect manner.”

History has already been made, with AlphaGo drawing first blood in the tournament of five matches. The ace landmark, of course, will be if the algorithm manages to win the series against a human mind that is currently ranked second in the number of international titles won in the game. The rest of the games will be played tomorrow, Saturday, Sunday and Tuesday, in Seoul local time (+3:30 IST). For a re-visit of today’s match, watch the full game here below: