A program developed by Google managed to defeat Lee Se-dol, the ancient of the “go play world champion”. The countless possible variations makes the game difficult to be encapsulated in fixed patterns, and much more based on the intuition of other chess boards
Share
a program developed by Google managed to defeat Lee Se -dol, the world champion of the ancient game of “go”. It is a major breakthrough for a new artificial intelligence model, which has enhanced intuitive abilities. The AlphaGo program managed to beat for the third time Lee, considered one of the strongest players of the modern era with 18 international titles under his belt, in a tournament made up of five games. For the moment, the human sample remained fasting.
The Go is an ancient Chinese game, very popular throughout East Asia and particularly in Japan, where the two players try to take control of the greater part of the board (goban). The countless possible variations makes the game difficult to be encapsulated in fixed patterns, and much more based on the intuition of other chess boards.
Since the historic victory of IBM’s Deep Blue computer on the then champion world Garry Kasparov in 1996, in chess computer programs are widely successful humans. But the go is a game far more complex and open to variants, for which until now you thought it could not be conquered by the machine. The same Lee until the eve of the challenge he had said he was convinced he could beat AlphaGo, although gradually scaring her was eased when she learned about the characteristics of its artificial challenger.
The Google team has developed the program using the most advanced artificial intelligence methods, based on neural networks with millions of connections. Its characteristic is to be able to eliminate a number of choices that humans instinctively consider them stupid.
No comments:
Post a Comment