Google’s AlphaZero AI Beats World’s Best Chess Software
AlphaZero artificial intelligence software triumphed at chess against the world’s best chess program after amassing all human knowledge of the game in just four hours.
Google researchers have published a paper in which they detail how their latest AI evolution, AlphaZero, managed to display a “superhuman performance” in chess, taking just four hours to learn the rules before destroying the hitherto ‘unbeatable’ world champion chess program, Stockfish.
The news comes a little more than a year after AlphaGo sensationally won against the world’s top Go player, a feat that never before been achieved.
Stockfish, which has the respect of the upper echelons of world chess players, and also won the 2016 TCEC Championship, and the 2017 Chess.com Computer Chess Championship, put up a brave performance but seemed utterly outclassed against Google’s advanced AI.
AlphaZero AI beats champion chess program after teaching itself in four hours
The results said it all. AlphaZero won the privately played 100-game match with 28 wins, 72 draws, and crucially, 0 losses. The 0 losses part is apparently the important bit.
Not the first
The open source project has though already been beaten by another program, Komodo, in two major computer chess challenges this year. It is perhaps hardly surprising. Stockfish was first released in 2008 and has been built on by volunteers in the years since, whereas AlphaZero has had the resources of the world’s largest tech company behind it.
“Starting from random play, and given no domain knowledge except the game rules, AlphaZero achieved within 24 hours a superhuman level of play in the games of chess and shogi [a similar Japanese board game] as well as Go, and convincingly defeated a world-champion program in each case,” said the paper’s authors, that include DeepMind founder Demis Hassabis, who was himself a child chess prodigy reaching the standard of ‘Master’ of the tender age of 13.
Past master
Former world champion Garry Kasparov told Chess.com that the result was hardly surprising, but no less of an achievement. “It’s a remarkable achievement, even if we should have expected it… It approaches the ‘Type B,’ human-like approach to machine chess dreamt of by Claude Shannon and Alan Turing instead of brute force.”
Reach for the stars
Google itself is not commenting on the research in an official capacity until the results have been peer reviewed. The London based team is however preparing for its next bigger, and even more complex challenge; trying to develop a system that can beat humans at the space strategy game Starcraft.
0 comments:
Post a Comment