What is the «credit assignment problem» in reinforcement learning?

See also: