Safe Reinforcement Learning by Imagining the Near Future

Open in new window