Meet Google DeepMind's AlphaGo, A Program That Beats Professional Player at Go
Once again, artificial intelligence (AI) researchers have proved that a computer learning program can perform better than humans when it comes to solve a complex puzzle. On Wednesday, DeepMind, a research organization under Google’ Alphabet, announced that its computer program has won against a high-ranking professional Go player.
The organization reported that the program, dubbed as AlphaGo, which combines two algorithms, defeated the human player in a series of five matches of highly complex Chinese strategy game. Results of the game have been published by journal Nature in its January 27 edition.
AI researchers consider Go, a traditional Chinese game which is played with black and white stones, as a good way to test someone’s complex problem solving skills. The strategic game is more complex and challenging than chess as there is far larger range of possible positions to make.
Before the latest competition between DeepMind’s program and professional human player, no software program had beaten humans in the game. The organization said in a statement that AlphaGo’s winning rate against other Go programs was over 99%. The program also defeated professional Go player from Europe, Fan Hui, DeepMind added.
The European Go champion played a series of five matches against AlphaGo in October last year. Since then, founder of DeepMind Technologies Demis Hassabis and other researchers associated with the program have been testing AlphaGo to make it better. DeepMind Technologies is a company acquired by Google more than one year ago.
About the program, Hassabis said, “The machine has continued to get better. We haven’t hit any kind of ceiling yet on performance”. He also said that designing an AI is a human endeavor and a kind of ingenious one, too.