The complexity of non-stationary reinforcement learning

Christos Papadimitriou, Binghui Peng

The problem of continual learning in the domain of reinforcement learning, often called non-stationary reinforcement learning, has been identified as an important challenge to the…