A reinforcement learning setting where the agent receives feedback infrequently, making learning difficult.