Autonomous Inter-Task Transfer In Reinforcement Learning Domains. door Matthew Edmund Taylor