Google DeepMind AlphaZero AI Annihilates Acclaimed Stockfish Chess Program In 100-Game Match

The greatest chess matches are no longer played by mere mortals comprised of flesh, bone, and blood, but sophisticated artificial intelligence (AI) schemes. Sitting at the top, at least for now, is AlphaZero, a product of Google's DeepMind division where machine learning rules the day. It also rules the chess board—AlphaZero took down Stockfish, the sophisticated open-source chess engine that many players use to prepare for big matches.

Stockfish's recent string of successes includes winning the 2016 TCEC Championship and and the 2017 Chess.com Computer Chess Championship. It is a formidable opponent by any standards, except when going up against AlphaZero. In a closed-door event, AlphaZero defeated Stockfish in 28 out of 100 games, tying the other 72 matches, ending the bout with zero losses. And this was after "learning" chess in just four hours.

What makes the feat even more remarkable is that AlphaZero used a type of machine learning to master the game of chess. it was not taught the game with traditional programming—there were no endgame tables inputted into AlphaZero, and no fancy algorithms to lean on. As it applies to machine learning, AlphaZero essentially taught itself how to win at chess, with just four hours to play against itself.

"It's a remarkable achievement, even if we should have expected it after AlphaGo," chess Grand Master Garry Kasparov told Chess.com. "It approaches the 'Type B,' human-like approach to machine chess dreamt of by Claude Shannon and Alan Turing instead of brute force."

The reference to AlphaGo is another DeepMind project, one that was able to defeat the world's top Go player, Ke Jie. AlphaGo also became a master by playing against itself instead of using algorithms. Both AlphaGo and AlphaZero are credits to DeepMind and the advances in machine learning that Google has made. This sort of thing also extends beyond board games.

"We have always assumed that chess required too much empirical knowledge for a machine to play so well from scratch, with no human knowledge added at all," Kasparov added. "Of course I’ll be fascinated to see what we can learn about chess from AlphaZero, since that is the great promise of machine learning in general—machines figuring out rules that humans cannot detect. But obviously the implications are wonderful far beyond chess and other games. The ability of a machine to replicate and surpass centuries of human knowledge in complex closed systems is a world-changing tool."

Indeed that's true. Machine learning has been used to help predict heart attacks, among other things, and will continue to be a point of focus by researchers and engineers.