DECAF: Deep Case-based Policy Inference for knowledge transfer in Reinforcement Learning

作者：

Highlights：

• Case-based Reasoning cycle in Reinforcement Learning terminology

• Algorithm for systematic retaining and reusing knowledge for complex tasks

• Building a library of core policies speeds up learning of new tasks

• Systematic approach to knowledge transfer avoids negative transfer

• Using favorable knowledge for transfer makes a big difference

摘要

•Case-based Reasoning cycle in Reinforcement Learning terminology•Algorithm for systematic retaining and reusing knowledge for complex tasks•Building a library of core policies speeds up learning of new tasks•Systematic approach to knowledge transfer avoids negative transfer•Using favorable knowledge for transfer makes a big difference

论文关键词：Deep Reinforcement Learning,Case-based Reasoning,Transfer Learning,Knowledge discovery,Knowledge management,Neural networks

论文评审过程：Received 20 May 2019, Revised 29 February 2020, Accepted 26 March 2020, Available online 31 March 2020, Version of Record 21 April 2020.

论文官网地址：https://doi.org/10.1016/j.eswa.2020.113420