Safety through Permissibility: Shield Construction for Fast and Safe Reinforcement Learning

Open in new window