DRL

EUCLID: Towards Efficient Unsupervised Reinforcement Learning with Multi-choice Dynamics Model

Figcps: Effective failure-inducing input generation for cyber-physical systems with deep reinforcement learning