A computer program to defeat first professional “go” – Terra Peru

The US company Google DeepMind has created a computer program that has been defeated for the first time, a professional oriental strategy game “go”, reveals a study published today in the British journal “Nature”.

One of the great challenges of artificial intelligence has been to achieve a win against a professional “go”, whose success, among other factors, is the ability to evaluate positions Player and movements on a board.

The new computer program developed by Google DeepMind, dubbed “AlphaGo” recently defeated in a game five games Fan Hui Chinese professional three-time European champion “go “- who failed to score a single sleeve in this clash against the machine.

Until now, computer programs “go” most successful came to play at the level of fans, but had never won a professional of the “oriental chess”.

The appearance of “AlphaGo” emphasize its creators, could allow significant progress in artificial intelligence domains, for now, seem inaccessible to researchers.

According to experts of Google DeepMind, “AlphaGo” use “value networks” to evaluate board positions and “tactical networks” to select the movement of the chips.

Artificial neural networks “AlphaGo” are trained in this game through a learning process overseen by a human being who, in turn, is reinforced by learning You get the machine to play against itself.

In games against other programs “go”, “AlphaGo” has won 99.8% of shocks, including held against professional Fan Hui, the researchers emphasize.

The British Go Association has issued a statement to congratulate Google DeepMind for this achievement, while waiting for the revenge in a match to be held in Seoul in March against the number one world during the past decade, South Korean Lee Sedol.

The popular game of “go” on a board facing two opponents that should alternately placing black and white stones on the free intersections of a grid of 19×19 lines.

The aim is to dominate the largest area of ​​the board, for which players must choose between placing the tiles together and protect each other to avoid being captured or separated so that they can gain ground in the grid.

Finding a balance between the two alternatives is the strategic difficulty of the ‘oriental chess’, appeared in China 3,000 years ago, and with a clear influence of Taoist spiritual principles.