The most sharp player in the classic strategic game Go is not a person ---it's a super-smart computer.
Google researchers
Alphmind has developed AlphaGo, an artificial intelligence system that mastered the art of ancient Chinese board games.
The team published their findings about computer success in nature.
Go is often regarded as the biggest challenge for artificial intelligence ---
Even more than chess. (
In 1997, IBM's Deep Blue, known for beating world chess champion Gary Kasparov, ushered in a new era of respect for computer brainpower. )
Popular live news: Tony Award's deadly crane collapse trophy hunting documentary Women's World Cup Go is a sophisticated game that includes a quick assessment of board position and a lot of potential moves.
"Until recently, the strongest computer was still playing the human amateur level of go," the researchers wrote . ".
How did you play?
It consists of two players alternately placing white and black pieces on a square board with a grid on it ---
Similar to a checkers or board, but with more space.
The ultimate goal is to fill more areas on the grid than your opponent.
Developers say the system, known as AlphaGo, has been successful where computers have failed in the past because it uses the "value network" to evaluate board positions and uses the "Policy Network"
AlphaGo "trained by the combination of supervised learning of human expert games and intensive learning of the game itself" is a process that simulates complex human thinking and decision-making --
They explained in a press release.
This kind of human cognitionlevel decision-
Production is "learning" from the trial. and-
Error, when the computer is against itself, it gets better every time.
AlphaGo scored 99 points in the test.
The artificial intelligence system has a "win rate" of up to 8% with other computer programs, eventually beating the European Go champion in one game.
Final result: AlphaGo won five games in the face of this daunting human opponent, marking the first time a computer program has won a pro.
Now researchers hope to test AlphaGo further.
The computing system will play the role of Li Shishi, a celebrity in the chess world.
Sedol is considered the world's number one player in the past decade.
The ultimate man-machine showdown will be held in Seoul this year.