PMID- 35249725 OWN - NLM STAT- MEDLINE DCOM- 20221012 LR - 20221012 IS - 1879-2022 (Electronic) IS - 0019-0578 (Linking) VI - 129 IP - Pt B DP - 2022 Oct TI - Model-driven deep unrolling: Towards interpretable deep learning against noise attacks for intelligent fault diagnosis. PG - 644-662 LID - S0019-0578(22)00087-8 [pii] LID - 10.1016/j.isatra.2022.02.027 [doi] AB - Intelligent fault diagnosis (IFD) has experienced tremendous progress owing to a great deal to deep learning (DL)-based methods over the decades. However, the "black box" nature of DL-based methods still seriously hinders wide applications in industry, especially in aero-engine IFD, and how to interpret the learned features is still a challenging problem. Furthermore, IFD based on vibration signals is often affected by the heavy noise, leading to a big drop in accuracy. To address these two problems, we develop a model-driven deep unrolling method to achieve ante-hoc interpretability, whose core is to unroll a corresponding optimization algorithm of a predefined model into a neural network, which is naturally interpretable and robust to noise attacks. Motivated by the recent multi-layer sparse coding (ML-SC) model, we herein propose to solve a general sparse coding (GSC) problem across different layers and deduce the corresponding layered GSC (LGSC) algorithm. Based on the ideology of deep unrolling, the proposed algorithm is unfolded into LGSC-Net, whose relationship with the convolutional neural network (CNN) is also discussed in depth. The effectiveness of the proposed model is verified by an aero-engine bevel gear fault experiment and a helical gear fault experiment with three kinds of adversarial noise attacks. The interpretability is also discussed from the perspective of the core of model-driven deep unrolling and its inductive reconstruction property. CI - Copyright (c) 2022 ISA. Published by Elsevier Ltd. All rights reserved. FAU - Zhao, Zhibin AU - Zhao Z AD - Xi'an Jiaotong University, PR China; School of Mechanical Engineering, Xi'an Jiaotong University, Xi'an, 710049, PR China; State Key Laboratory for Manufacturing Systems Engineering, Xi'an Jiaotong University, Xi'an, 710049, PR China. Electronic address: zhaozhibin@xjtu.edu.cn. FAU - Li, Tianfu AU - Li T AD - Xi'an Jiaotong University, PR China; School of Mechanical Engineering, Xi'an Jiaotong University, Xi'an, 710049, PR China; State Key Laboratory for Manufacturing Systems Engineering, Xi'an Jiaotong University, Xi'an, 710049, PR China. Electronic address: litianfu@stu.xjtu.edu.cn. FAU - An, Botao AU - An B AD - Xi'an Jiaotong University, PR China; School of Mechanical Engineering, Xi'an Jiaotong University, Xi'an, 710049, PR China; State Key Laboratory for Manufacturing Systems Engineering, Xi'an Jiaotong University, Xi'an, 710049, PR China. Electronic address: albert_an@stu.xjtu.edu.cn. FAU - Wang, Shibin AU - Wang S AD - Xi'an Jiaotong University, PR China; School of Mechanical Engineering, Xi'an Jiaotong University, Xi'an, 710049, PR China; State Key Laboratory for Manufacturing Systems Engineering, Xi'an Jiaotong University, Xi'an, 710049, PR China. Electronic address: wangshibin2008@xjtu.edu.cn. FAU - Ding, Baoqing AU - Ding B AD - Xi'an Jiaotong University, PR China; School of Mechanical Engineering, Xi'an Jiaotong University, Xi'an, 710049, PR China; State Key Laboratory for Manufacturing Systems Engineering, Xi'an Jiaotong University, Xi'an, 710049, PR China. Electronic address: dingbq@xjtu.edu.cn. FAU - Yan, Ruqiang AU - Yan R AD - Xi'an Jiaotong University, PR China; School of Mechanical Engineering, Xi'an Jiaotong University, Xi'an, 710049, PR China; State Key Laboratory for Manufacturing Systems Engineering, Xi'an Jiaotong University, Xi'an, 710049, PR China. Electronic address: yanruqiang@xjtu.edu.cn. FAU - Chen, Xuefeng AU - Chen X AD - Xi'an Jiaotong University, PR China; School of Mechanical Engineering, Xi'an Jiaotong University, Xi'an, 710049, PR China; State Key Laboratory for Manufacturing Systems Engineering, Xi'an Jiaotong University, Xi'an, 710049, PR China. Electronic address: chenxf@xjtu.edu.cn. LA - eng PT - Journal Article DEP - 20220222 PL - United States TA - ISA Trans JT - ISA transactions JID - 0374750 SB - IM MH - Algorithms MH - *Deep Learning MH - Neural Networks, Computer OTO - NOTNLM OT - Intelligent fault diagnosis OT - Interpretable deep learning OT - Model-driven deep unrolling OT - Noise attacks COIS- Declaration of Competing Interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. EDAT- 2022/03/08 06:00 MHDA- 2022/10/13 06:00 CRDT- 2022/03/07 05:38 PHST- 2021/01/16 00:00 [received] PHST- 2021/12/26 00:00 [revised] PHST- 2022/02/14 00:00 [accepted] PHST- 2022/03/08 06:00 [pubmed] PHST- 2022/10/13 06:00 [medline] PHST- 2022/03/07 05:38 [entrez] AID - S0019-0578(22)00087-8 [pii] AID - 10.1016/j.isatra.2022.02.027 [doi] PST - ppublish SO - ISA Trans. 2022 Oct;129(Pt B):644-662. doi: 10.1016/j.isatra.2022.02.027. Epub 2022 Feb 22.