Simultaneously Learning Stochastic and Adversarial Episodic MDPs with Known Transition

Aug-16-2025, 04:55:16 GMT–Neural Information Processing Systems 

We study the problem of learning episodic Markov Decision Processes (MDPs).