Generalization in reinforcement learning: Safely approximating the value function (1995)

by Justin A Boyan, Andrew W Moore
Venue:Advances in Neural Information Processing Systems 7