Deep transfer learning for conditional shift in regression

作者：

Highlights：

• This paper reveals a general type of conditional shift regression problem in DTL.

• A metric CEOD is proposed to measure the conditional distribution discrepancy.

• The global properties of conditional distribution of the target data are considered.

摘要

•This paper reveals a general type of conditional shift regression problem in DTL.•A metric CEOD is proposed to measure the conditional distribution discrepancy.•The global properties of conditional distribution of the target data are considered.

论文关键词：Deep transfer learning,Conditional shift,Regression,Kernel embedding

论文评审过程：Received 9 February 2021, Revised 11 May 2021, Accepted 9 June 2021, Available online 11 June 2021, Version of Record 20 June 2021.

论文官网地址：https://doi.org/10.1016/j.knosys.2021.107216