Deep multi-scale separable convolutional network with triple attention mechanism: A novel multi-task domain adaptation method for intelligent fault diagnosis

作者：

Highlights：

• A novel unsupervised domain adaptation network, termed MSSCN-TAM, is established.

• Triple attention mechanism is constructed to enhance the self-adjusting ability.

• The MSSCN-TAM alleviates the dependence on the label information in target domain.

• The MSSCN-TAM has superior adaptability and transferability for the variable domains.

摘要

•A novel unsupervised domain adaptation network, termed MSSCN-TAM, is established.•Triple attention mechanism is constructed to enhance the self-adjusting ability.•The MSSCN-TAM alleviates the dependence on the label information in target domain.•The MSSCN-TAM has superior adaptability and transferability for the variable domains.

论文关键词：Fault diagnosis,Domain adaptation,Multi-scale separable convolution network,Maximum mean discrepancies,Attention mechanism

论文评审过程：Received 7 November 2020, Revised 5 March 2021, Accepted 17 April 2021, Available online 23 May 2021, Version of Record 5 June 2021.

论文官网地址：https://doi.org/10.1016/j.eswa.2021.115087