Preferential Temporal Difference Learning

Nishanth V. Anand, Doina Precup

Abstract: Temporal-Difference (TD) learning is a general and very useful tool for estimating the value function of a given policy, which in turn is required to find good policies. Generally speaking, TD lear...