A computer has beaten a human champion at the ancient games of Go. This is a new milestone for artificial intelligence able to win at a strategy game that rather than needing processing power requires intuition, a human quality.
The artificial intelligence machine is called AlphaGo. The system learned its Go skills through a process of trial and error, by playing against itself millions of games. Now it was able to defeat a Go champion with five-nil, surprising even its creators. Demis Hassabis, the chief executive of Google DeepMind, a British artificial intelligence (AI) company acquired by Google, declared that AphaGo was stronger than the developer team expected.
The ancient Chinese game of Go is one of the most complex ever designed. According to the website phys.org, Hassabis said that Go is a game with many configurations possible and the AphaGo machine is a major step forward in the development of artificial intelligence. The British Go Association declared in a statement that the victory over three-time European Go champion Fan Hui was unexpected.
In the game of Go, two players take their turn and place stones on a board. The aim of the game is to surround and capture the opponent's stones, trying to control more than 50 percent of the board. There are many places where a player can place the first stone and many ways in which the opponent can respond to these moves.
Despite the fact that the game has simple rules, it provides profound complexity. Such complexity is too vast for brute processing power approach, according to Hassabis' colleague David Silver, co-author of the paper published in the science journal Nature. For this reason, the artificial intelligence developer team aimed to create a machine with a "more human-like" approach that simulates human intuition.
DeepMind's AlphaGo features millions of neuron-like connections organized in two sets of "deep neural networks". This allows the machine to reduce the search base and look ahead to the new moves in a way akin to intuition.
