Weakly-Supervised ReinforcementLearningfor ControllableBehavior

Neural Information Processing Systems 

We show that thislearned subspace enables efficient exploration andprovides arepresentation that captures distance between states.