GAIL

Generative Adversarial Imitation Learning (GAIL) is a framework that combines generative adversarial networks with imitation learning to enable agents to learn from expert data without explicit reward functions.