Importance-weighted conditional adversarial network for unsupervised domain adaptation

作者：

Highlights：

• Propose a deep adversarial adaptation network for unsupervised domain adaptation(DA).

• The network contributes to reducing the harmful impact of hard-to-transfer samples.

• Derive a new sample selection criterion to improve target domain discriminability.

• Extensive experimentation shows its effectiveness over previous DA methods.

摘要

•Propose a deep adversarial adaptation network for unsupervised domain adaptation(DA).•The network contributes to reducing the harmful impact of hard-to-transfer samples.•Derive a new sample selection criterion to improve target domain discriminability.•Extensive experimentation shows its effectiveness over previous DA methods.

论文关键词：Domain adaptation,Deep learning,Adversarial learning,Importance weightage,Sample selection

论文评审过程：Received 19 August 2019, Revised 20 March 2020, Accepted 23 March 2020, Available online 4 April 2020, Version of Record 26 April 2020.

论文官网地址：https://doi.org/10.1016/j.eswa.2020.113404