Unsupervised Domain Adaptation via Deep Conditional Adaptation Network

作者：

Highlights：

• A novel domain adaptation (DA) framework is proposed by aligning the conditional distributions and simultaneously extracting discriminant information from both domains.

• Conditional maximum mean discrepancy is used to align the conditional distributions directly by their conditional embeddings in reproducing kernel Hilbert space.

• We further extend the proposed DA framework on partial DA scenarios.

• The proposed method achieves state-of-the-art performance on both DA and partial DA scenarios.

摘要

•A novel domain adaptation (DA) framework is proposed by aligning the conditional distributions and simultaneously extracting discriminant information from both domains.•Conditional maximum mean discrepancy is used to align the conditional distributions directly by their conditional embeddings in reproducing kernel Hilbert space.•We further extend the proposed DA framework on partial DA scenarios.•The proposed method achieves state-of-the-art performance on both DA and partial DA scenarios.

论文关键词：Deep learning,Domain adaptation,Feature extraction,Conditional maximum mean discrepancy,Kernel method

论文评审过程：Received 18 May 2022, Revised 26 August 2022, Accepted 28 September 2022, Available online 1 October 2022, Version of Record 14 October 2022.

论文官网地址：https://doi.org/10.1016/j.patcog.2022.109088