Cross-domain fault diagnosis of bearing using improved semi-supervised meta-learning towards interference of out-of-distribution samples
作者：
Highlights：
• A new improved semi-supervised meta-learning model is proposed.
• LAS is designed to fully exploit the unlabelled information.
• Scalable distance metric function is defined.
• The proposed method can effectively suppress the OOD interference samples.
• Different cross-domain scenarios and OOD samples are set for verification.
摘要
•A new improved semi-supervised meta-learning model is proposed.•LAS is designed to fully exploit the unlabelled information.•Scalable distance metric function is defined.•The proposed method can effectively suppress the OOD interference samples.•Different cross-domain scenarios and OOD samples are set for verification.
论文关键词：Cross-domain bearing fault diagnosis,Out-of-distribution samples,Label allocation strategy,Scalable distance metric function,Meta-learning
论文评审过程：Received 20 May 2022, Revised 5 July 2022, Accepted 16 July 2022, Available online 22 July 2022, Version of Record 2 August 2022.
论文官网地址：https://doi.org/10.1016/j.knosys.2022.109493
