As a critical component in agriculture, industry, and the military, pump anomaly detection has recently aroused wide attention, which requires deep and abundant development and application. Researchers emphasize deeper networks that are long vast computational resources despite insufficient training samples prepared. To break through this obstacle above, we propose a few-shot model-agnostic meta-learning strategy (MAMLS) model to mitigate the data scarcity problem. Inspired by the diffusive ordinary differential equations (ODEs) and Wide-Resnet (WRN), we made great strides by connecting diffusion (Diff) mechanism and self-adaptive Lr with MAMLS. We generate two classical synthetic datasets (circle and spiral) to clarify the diffusion algorithm’s capability to enhance the relationships and weaken the noise. The experimental results under synthetic data confirm that accuracy quickly reached 99% after several iterations. In an actual case anomaly detection study on the pumps simulation platform, the proposed Diff-WRN-MAMLS brings substantial advantages in saving hardware resources. Compared to current models, our model achieves 98% accuracy in 9-way 25-shot tasks. In the operating efficiency experiment, our algorithm only consumed 14.37 quality factors. The final experiment with four state-of-the-art model-agnostic meta-learning (MAML)-enhanced methods demonstrates the highest reliable test accuracy in different cases, reaching 98.5, 97.8, and 98.4%, respectively. Results showed that the proposed method will generalize surprisingly well in anomaly detection in future research.
WangCYGaoJZLiH, et al. Robust sparse Bayesian learning for broad learning with application to high-speed railway track monitoring. Struct Health Monit Int J. Epub ahead of print 10 June 2022. DOI: 10.1177/14759217221104224.
2.
WangWTLiHWangCH, et al. An improved ultrasonic computerized tomography (UCT) technique for damage localization based on compressive sampling (CS) theory. Struct Control Health Monit2022; 29(6): e2938.
3.
LiXLiangYZhaoM, et al. Few-shot learning with generative adversarial networks based on WOA13 data. CMC-Comput Mater Continua2019; 60(3): 1073–1085.
4.
TnaniM-ASubarnadutiPDiepoldK.Efficient feature learning approach for raw industrial vibration data using two-stage learning framework. Sensors2022; 22(13): 4813.
5.
ZhouXLiangWShimizuS, et al. Siamese neural network based few-shot learning for anomaly detection in industrial cyber-physical systems. IEEE Trans Ind Inf2021; 17(8): 5790–5798.
6.
FangQWuD.ANS-net: anti-noise Siamese network for bearing fault diagnosis with a few data. Nonlinear Dyn2021; 104: 2497–2514.
7.
LuJJinSLiangJ, et al. Robust few-shot learning for user-provided data. IEEE Trans Neural Networks Learn Syst2021; 32(4): 1433–1447.
SuJShenHPengL, et al. Few-shot domain-adaptive anomaly detection for cross-site brain images. IEEE Trans Pattern Anal Mach Intell2021; PP(99): 1.
10.
TakimotoHSekiJSitujuSF, et al. Anomaly detection using siamese network with attention mechanism for few-shot learning. Appl Artif Intell2022; 36(1).
11.
YeTLiGAhmadI, et al. FLAG: few-shot latent dirichlet generative learning for semantic-aware traffic detection. IEEE Trans Netw Serv Manage2022; 19(1): 73–88.
12.
BaoYLiH.Machine learning paradigm for structural health monitoring. Struct Health Monit2020; 20(4): 1353–1372.
13.
LiJQiuXWeiY, et al. Online rail fastener detection based on YOLO network. CMC-Comput Mater Continua2022; 72(3): 5955–5967.
14.
ZhangALiSCuiY, et al. Limited data rolling bearing fault diagnosis with few-shot learning. IEEE Access2019; 7: 110895–110904.
15.
XuYLiYWangY, et al. Improved few-shot learning method for transformer fault diagnosis based on approximation space and belief functions. Expert Syst Appl2021; 167(1): 114105.
16.
LiuCQinCShiX, et al. TScatNet: an interpretable cross-domain intelligent diagnosis model with antinoise and few-shot learning capability. IEEE Trans Instrum Meas2021; 70: 1–10.
17.
AhmadZNguyenTKAhmadS, et al. Multistage centrifugal pump fault diagnosis using informative ratio principal component analysis. Sensors (Basel)2021; 22(1): 179.
18.
VigneswaranRLawMTBalasubramanianVN, et al. Feature generation for long-tail classification. In: Proceedings of the twelfth Indian conference on computer vision, graphics and image processing, 2021, New York, NY, pp. 1–9.
19.
ZeleninaLKhaiminaLKhaiminE, et al. Convolutional neural networks in the task of image classification. Math Inform2022; 65: 19–29.
20.
ZhuJJiangQShenY, et al. Application of recurrent neural network to mechanical fault diagnosis: a review. J Mech Sci Technol2022; 36: 527–542.
21.
PeiZJiangHLiX, et al. Data augmentation for rolling bearing fault diagnosis using an enhanced few-shot Wasserstein auto-encoder with meta-learning. Meas Sci Technol2021; 32: 084007.
22.
LiuSChenJQuC, et al. LOSGAN: latent optimized stable GAN for intelligent fault diagnosis with limited data in rotating machinery. Meas Sci Technol2021; 32(4): 045101.
23.
SatoKNakataSMatsubaraT, et al. Few-shot anomaly detection using deep generative models for grouped data. IEICE Trans Inf Syst2022; E105D: 436–440.
24.
WangDZhangMXuY, et al. Metric-based meta-learning model for few-shot fault diagnosis under multiple limited data conditions. Mech Syst Sig Process2021; 155: 107510.
25.
LiCLiSZhangA, et al. Meta-learning for few-shot bearing fault diagnosis under complex working conditions. Neurocomputing2021; 439(2): 197–211.
26.
GuoZZhangCYuW, et al. Few-shot graph learning for molecular property prediction. In: 30th World Wide Web Conference (WWW), 2021, pp. 2559–2567. ACM.
27.
WangYJiangYLanJ.Intrusion Detection Using Few-shot Learning Based on Triplet Graph Convolutional Network. J Web Eng2021; 20(5): 1527–1552.
28.
FengYChenJXieJ, et al. Meta-learning as a promising approach for few-shot cross-domain fault diagnosis: algorithms, applications, and prospects. Knowledge Based Syst2022; 235: 107646.
29.
ZhangSYeFWangB, et al. Few-shot bearing anomaly detection via model-agnostic meta-learning. In: 23rd International Conference on Electrical Machines and Systems (ICEMS), 2020, pp. 1341–1346. ACM.
30.
DoraVLakshmiVN.Optimal feature selection with CNN-feature learning for DDoS attack detection using meta-heuristic-based LSTM. Int J Intell Rob Appl2022; 6(2): 1–27.
31.
ChuXHWangJYLiSX, et al. Empirical study on meta-feature characterization for multi-objective optimization problems. Neural Comput Appl2022; 34(19): 16255–16273.
32.
FinnCAbbeelPLevineS.Model-agnostic meta-learning for fast adaptation of deep networks. In: Proceedings of the 34th international conference on machine learning, Sydney, Australia, 2017.
33.
KalauchAPicardRSiegmundS, et al., A hilbert space perspective on ordinary differential equations with memory term. J Dyn Differ Equations2014; 26: 369–399.
34.
ZagoruykoSKomodakisN.Wide residual networks. In: British machine vision conference 2016, University of York, York, UK2016.
35.
LiuJWWangYFLuRK, et al. Multi-view non-negative matrix factorization discriminant learning via cross entropy loss. In: 32nd Chinese control and decision conference (CCDC), Hefei, China, 2020, pp. 3964–3971.
36.
PalMRaikarAPandaA, et al. Synthetic speech detection using meta-learning with prototypical loss. 2022.
37.
LuNYinT.Transferable common feature space mining for fault diagnosis with imbalanced data. Mech Syst Sig Process2021; 156: 107645.
38.
WangTDouZBaoC, et al. Diff-ResNets for Few-shot Learning – an ODE Perspective. 2021.
39.
PyoYoung-IlPark, et al. Noise reduction in high-ISO images using 3-D collaborative filtering and structure extraction from residual blocks. IEEE Trans Consum Electron2011; 57(2): 687–695.
40.
JainS.Hypothesis spaces for learning. Inf Comput2011; 209(3): 513–527.
JainVPalhaAGerritsmaM.Use of algebraic dual representations in domain decomposition methods for Darcy flow in 3D domain. 2022.
43.
GolubMLemieuxGLisM.DropBack: Continuous Pruning During Training. 2018.
44.
PanTChenJXieJ, et al. Deep feature generating network: a new method for intelligent fault detection of mechanical systems under class imbalance. IEEE Trans Ind Inf2020; PP(99): 1–1.
45.
VenskauskasKK. Optimization of signal-detection against a background of non-Gaussian noise. Telecommun Radio Eng1987; 41(2): 91–92.
46.
ZhuangYQTaoLYangF, et al. RelationNet: learning deep-aligned representation for semantic image segmentation. In: 24th International Conference on Pattern Recognition (ICPR), Chinese Acad Sci, Inst Automat, Beijing, Peoples Republic of China, 20–24 August 2018, pp. 1506–1511.
47.
VinyalsOBlundellCLillicrapT, et al. Matching networks for one shot learning. In: 30th conference on neural information processing systems (NIPS 2016), Barcelona, Spain, 2016.
48.
HuangX-LChoiSH.Intra-class local descriptor-based prototypical network for few-shot learning. J Korea Multimedia Society2022; 25(1): 52–60.
49.
WangQWuWZhaoY, et al. Graph active learning for GCN-based zero-shot classification. Neurocomputing2021; 435: 15–25.
50.
AhmedIQuinones-GrueiroMBiswasG. Complementary Meta-Reinforcement Learning for Fault-Adaptive Control. In: arXiv2020.
51.
LuYPengXYangD, et al. Model-agnostic meta-learning with optimal alternative scaling value and its application to industrial soft sensing. IEEE Trans Ind Inf2021;17(12): 8003–8013.
52.
DixitSVermaNKGhoshAK.Intelligent fault diagnosis of rotary machines: conditional auxiliary classifier GAN coupled with meta learning using limited data. IEEE Trans Instrum Meas2021; PP(99): 1.
53.
ZhangKTangBDengL, et al. A fault diagnosis method for wind turbines gearbox based on adaptive loss weighted meta-ResNet under noisy labels. Mech Syst Sig Process2021; 161: 107963.
