Describir: Model-based Reinforcement Learning with Model Error and Its Application