A Google computer defeat the world champion of an ancient game

[-INTOX-™]

 

The software AlphaGo achieves a milestone in artificial intelligence by winning the first game of Go against South Korean Lee Sedol. A million dollars are at stake.



The prestigious journal Nature announced with great fanfare last January a milestone of artificial intelligence that is not thought possible in a decade. A computer program developed by the British company Google DeepMind had managed to win for the first time, a professional champion of an ancient game called Go Eastern origin. The challenge was huge for a machine, since the test strategy is a complex concept. Now the software AlphaGo has outdone himself has won no less than the world champion Go, South Korean Lee Sedol. The game, held in a downtown hotel in Seoul this past dawn, was followed live by thousands of people around the world through internet.
Go is a test of strategy po[CENSORED]r in countries like China, South Korea or Japan. In it, two players must conquer territory as possible by placing a black and white stones on a board. The rules are simple, but the chances of multiple game because they depend on the creativity, intelligence, experience and style of player. Therefore, fans say it is more sophisticated than chess.
Until he came AlphaGo, the most successful computer programs developed for this game were unable to defeat a human professional player. But this software seems to them all. After making history by beating the European champions, the Chinese Fan Hui, now AlphaGo has overcome an even greater obstacle to emerge victorious from the first of a series of five games against Sedol, professional player of 32 years and 18 times winner international title. He did the 186 movements and after nearly three hours of departure. Although commentators described as "unusual" some of the movements of the program, it was doing with the board to corner the South Korean.
Sedol, recognized as the best Go player of the last decade worldwide, had shown before the game very confident in their own abilities, but then could not help but concede defeat to enthusiastic audiences.

One million dollars

You can still take revenge. The new challenge of man versus machine runs until next Tuesday 15. The winner of the series a prize of one million dollars will be. If you win the machine, Google has announced that it will donate the money to UNICEF.
The ability to AlphaGo to learn from each other in a human-like way it is the key to its success, unlike, for example, the IBM supercomputer Deep Blue, which clashed on several occasions between 1996 and 1997 champion Russian chess Garry Kasparov. The program uses different networks to evaluate positions on the board and select movements. It is what is called artificial neural networks that mimic biological. They were prepared by combining learning supervised by expert players with learning that is achieved when the machine plays against itself.
In recent years, scientists have made great strides in getting computers to think and learn in a similar way to the people, with the ultimate goal of artificial intelligence may someday help humans in areas such as health and scientific investigation.