Averaging n-step Returns Reduces Variance in Reinforcement Learning.

Brett Daley, Martha White, Marlos C. Machado

Browse the full ICML paper archive.