It almost 20 years ago IBM:s Deep Blue super-computer beat the reign world chess champion Gary Kasparov the first time. Deep Blue’s basic method to play the game was to evaluate every possible move, which would mean 200 million per second in DP:s case. While humans that plays at the same level, which is unable to do this, use a probability method that carry down the game to a few possible profitable roads to play. This is a method that Artificial Intelligence has had difficulties to use, until now. From this background Matthew Lai at Imperial College London has created an artificial intelligence machine called Giraffe. This machine is teaching itself to play chess by evaluating positions more like humans probability method. Since it is firstly focusing on evaluate the board rather than every possible move. Source: MITNews
Latest News: Next Generation Artificial Intelligence: Machine Teaches itself
Bookmark the permalink.