First, AlphaGo beat the best human player in the world by studying thousands of human vs. human games.

Then AlphaGo Zero came along and taught itself to be even better without any human generated data. By the way, it beat AlphaGo.

This is the power of Reinforcement Learning.

tt ads