Imitation Learning in Continuous Action Spaces: Mitigating Compounding Error without Interaction

Open in new window