Compressing Deep Reinforcement Learning Policies into Symbolic Expressions [pdf]

proceedings.mlr.press