Imitation Learning in Continuous Action Spaces: Mitigating Compounding Error without InteractionShare on Twitter Facebook LinkedIn Previous Next