AlphaZero is a DeepMind algorithm that taught itself chess, shogi and Go from scratch, given only the rules, and beat top programs in all three.
In plain words
AlphaGo first learned from human games. AlphaZero started with nothing but the rules and random moves. It improved by playing game after game against itself, using reinforcement learning to update a neural network that guides its search. The same general method worked for three different games, and within 24 hours of training it had reached superhuman strength in each one.
A quick example
At the start of training AlphaZero's chess moves are random. After many self-play games, its network has learned which positions tend to win, and it defeated a world-champion chess program.
Where you'll meet it
- Research on self-play and general game-playing AI.
Related
Sources
- PaperMastering Chess and Shogi by Self-Play with a General Reinforcement Learning Algorithm, Silver et al. (arXiv) · read 29 Sept 2026
- DocsAlphaGo Zero: Starting from scratch, Google DeepMind · read 29 Sept 2026
Checked by WhatIsAI against the sources above · 29 Sept 2026. Spotted an error? Report it.