Smooth Imitation Learning via Smooth Costs and Smooth Policies
Imitation learning (IL) is a popular approach in the continuous control setting as among other reasons it circumvents the problems of reward mis-specification and exploration in reinforcement learning (RL). In IL from …