Google says its DeepMind unit's AlphaGo Zero algorithm has defeated world-leading specialist software Stockfish 8 in a chess match within hours of teaching itself the game.
AlphaGo Zero reportedly became capable of outperforming Stockfish only four hours after being given the rules of chess and being instructed to learn by playing simulations against itself.
Each program was given 60 seconds' worth of thinking time for each move in the 100 matches that followed, and AlphaGo Zero won 28 games, while the remaining 72 ended in a draw.
"From a scientific point of view, it's the latest in a series of dazzling results that DeepMind has produced," says University of Oxford in the U.K. professor Michael Wooldridge, who received the ACM/SIGART Autonomous Agents Research Award for 2006. He notes while the games AlphaGo Zero is trained to play are "closed" in the sense that they have limited sets of rules, "things will get even more exciting when DeepMind moves on to more open problems."
From BBC News
View Full Article
Abstracts Copyright © 2017 Information Inc., Bethesda, Maryland, USA
No entries found