Model-Based Reinforcement Learning via Meta-Policy Optimization

摘要: … the asymptotic performance of model-free methods while … model-based RL and approaches that combine elements of … the real environment, and fine-tuning the policy with VPG on the …