论文部分内容阅读
This paper studies the problem of transfer learning in the context of reinforcement learning.We propose a novel transfer learning method that can speed up reinforcement learning with the aid of previously learnt tasks.Before performing extensive learning