Evolution-Guided Policy Gradient in Reinforcement Learning

Temporal Difference methods inRL use bootstrapping to address this issue but often struggle when the time horizons are long and the reward is sparse.