DeepMind, a company owned by Google that focuses on development in Artificial Intelligence, on Wednesday, designed a chess programme on the AlphaZero software that can take on and beat the strongest of opponents. It took all of 24 hours for the software to teach itself the rules game and master it.
The recent set of results have made experts sit up and take notice after AlphaZero after it played against the latest version of the Stockfish programme recently and won by a huge margin of 64:36, not lose a single game, Chessbase.com reported. The software also has a chess connection with Demis Hassabis, the founder of DeepMind was once ranked as the second highest U-14 player in the world.
“From a scientific point of view, it’s the latest in a series of dazzling results that DeepMind has produced,” the University of Oxford’s Prof Michael Wooldridge was quoted as saying by BBC. “The general trajectory in DeepMind seems to be to solve a problem and then demonstrate it can really ramp up performance, and that’s very impressive.”
It was only three years ago that DeepMind designed the ancient board game go. The game’s legend, South Korea’s Lee Sodol was comfortably beaten 4-1. The programme continued to earn the plaudits after an updated version AlphaGo beat Chinese prodigy Ke Jie 3-0.