Adversarial Skill Chaining for Long-Horizon Robot Manipulation via Terminal State Regularization (CoRL 2021) - View it on GitHub