ConceptSelf-Play / 自博弈
Self-Play
A training method in reinforcement learning in which an AI improves by competing against different versions of itself. This approach causes training difficulty to scale automatically with model capability and served as a core mechanism enabling game-playing AI systems such as AlphaGo to achieve superhuman performance.
Timeline (last 90 days)
Sep 10
OpenAI once trained agents capable of playing hide-and-seek through self-play, with complex behaviors such as tool use, shelter construction, and exploitation of physics exploits emerging spontaneously.
Unverified50%