Manifold discrimination partial adversarial domain adaptation

作者：

Highlights：

• MDPDA is proposed as a model framework. It designs a manifold discrimination sample weighting mechanism.

• The manifold alignment and graph based source domain sample similarity discrimination are constructed.

• We provide comprehensive experiments in three datasets to validate the good perfomance of our method.

摘要

•MDPDA is proposed as a model framework. It designs a manifold discrimination sample weighting mechanism.•The manifold alignment and graph based source domain sample similarity discrimination are constructed.•We provide comprehensive experiments in three datasets to validate the good perfomance of our method.

论文关键词：Partial domain adaptation,Domain adaptation,Transfer learning,Manifold learning,Image classification

论文评审过程：Received 22 June 2021, Revised 21 June 2022, Accepted 21 June 2022, Available online 25 June 2022, Version of Record 9 July 2022.

论文官网地址：https://doi.org/10.1016/j.knosys.2022.109320