Learning Robust Feature Transformation for Domain Adaptation

作者：

Highlights：

• We propose Robust Transfer Feature Learning (RTFL) for domain adaptation.

• An iterative optimization procedure is proposed with guaranteed convergence.

• Efficient RTFL is further developed, admitting low computation complexity.

• The methods achieve state-of-the-art results on some datasets under severe noises.

摘要

•We propose Robust Transfer Feature Learning (RTFL) for domain adaptation.•An iterative optimization procedure is proposed with guaranteed convergence.•Efficient RTFL is further developed, admitting low computation complexity.•The methods achieve state-of-the-art results on some datasets under severe noises.

论文关键词：Domain adaptation,Linear transformation,Correntropy,Kernel mean p-power error loss,Half-quadratic optimization

论文评审过程：Received 1 January 2019, Revised 30 November 2020, Accepted 31 January 2021, Available online 5 February 2021, Version of Record 11 February 2021.

论文官网地址：https://doi.org/10.1016/j.patcog.2021.107870