Divide and Repair: Using Options to Improve Performance of Imitation Learning Against Adversarial Demonstrations

Open in new window