Self-play

Reinforcement learning technique / From Wikipedia, the free encyclopedia

Self-play is a technique for improving the performance of reinforcement learning agents. Intuitively, agents learn to improve their performance by playing "against themselves".

Oops something went wrong: