Stacked Universal Successor Feature Approximators for Safety in Reinforcement Learning

Open in new window