Evolution of reward functions for reinforcement learning.

Scott Niekum, Lee Spector, Andrew G. Barto

Browse the full GECCO paper archive.