The program AlphaGo won the first of a series of games against one of the strongest masters of a go — Korean Lee Sedalia (second largest in the world by number of international titles). It is reported by Gizmodo.
The competition takes place from 9 to 13 March. Every day artificial intelligence and Whether SEDOL fight once. YouTube will broadcast games live.
According to the staff of DeepMind (developer AlphaGo), the program, despite the small errors, confidently beat the champion.
Logical Board game of go, which arose in Ancient China, is considered the “hard nut” for artificial intelligence — due to the large size of the Board and of the difficulties in predicting moves.
Play two: one gets the black stones, the other white. The goal of opponents is to isolate on the game Board with stones of his color more territory than the opponent. The best program has reached only to the level of Amateur players.
AlphaGo program developed by Google-owned DeepMind company, uses the so-called network values to assess the situation of the pieces on the Board and the network rules for the choice of moves. These neural networks are studying the game, analyzing well-known party as well as by trial and error, playing alone.
AlphaGo beat other programs in 99.8 percent of parties, and then achieved the superiority over European champion of the game of go (with a score of five to zero) without any handicap.
Related posts:
Artificial intelligence defeated the world champion in the game in th
Mature artificial intelligence we have already created based on self-organizing Smart-MES
Began live broadcasting of the game of the neural network against the world champion of the game in th
Artificial intelligence DeepMind Google never ceases to amaze
