Deep Transfer Learning With Self-Attention for Industry Sensor Fusion Tasks
Open Access
- 1 July 2022
- journal article
- research article
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Sensors Journal
- Vol. 22 (15) , 15235-15247
Abstract
Monitoring of complex industrial processes can be achieved by obtaining process data by utilising various sensing modalities. The recent emergence of deep learning provides a new routine for processing multi-sensor information. However, the learning ability of shallow neural networks is insufficient, and the data amount required by deep networks is often too large for industrial scenarios. This paper provides a novel deep transfer learning method as a possible solution that offers an advantage of better learning ability of the deep network without the requirement for a large amount of training data. This paper presents how Transformer with self-attention trained from natural language can be transferred to the sensor fusion task. Our proposed method is tested on 3 datasets: condition monitoring of a hydraulic system, bearing, and gearbox dataset. The results show that the Transformer trained from natural language can effectively reduce the required data amount for using deep learning in industrial sensor fusion with high prediction accuracy. The difficult and uncertain artificial feature engineering which requires a large workload can also be eliminated, as the deep networks are able to extract features automatically. In addition, the self-attention mechanism of Transformer aids in the identification of critical sensors, hence the interpretability of deep learning in industrial sensor fusion can be improved.Keywords
Funding Information
- EPSRC through the Future Electrical Machines Manufacturing Hub (EP/S018034/1)
- UK Research and Innovation/Engineering and Physical Sciences Research Council (UKRI/EPSRC) through the Made Smarter Innovation-Research Centre for Connected Factories (EP/V062123/1)
- Royal Academy of Engineering (RAEng) and Airbus through the Research Chairs and Senior Research Fellowships Scheme (RCSRF1718/5/41)
This publication has 36 references indexed in Scilit:
- Machine Health Monitoring Using Local Feature-Based Gated Recurrent Unit NetworksIEEE Transactions on Industrial Electronics, 2017
- Energy-Fluctuated Multiscale Feature Learning With Deep ConvNet for Intelligent Spindle Bearing Fault DiagnosisIEEE Transactions on Instrumentation and Measurement, 2017
- Coil winding process modelling with deformation based wire tension analysisCIRP Annals, 2016
- Application of Clustering Methods for Online Tool Condition Monitoring and Fault Diagnosis in High-Speed Milling ProcessesIEEE Systems Journal, 2015
- Condition monitoring of a complex hydraulic system using multivariate statistics2015
- Effective Approaches to Attention-based Neural Machine Translation2015
- Wavelet leaders multifractal features based fault diagnosis of rotating mechanismMechanical Systems and Signal Processing, 2014
- Motor Bearing Fault Diagnosis Using Trace Ratio Linear Discriminant AnalysisIEEE Transactions on Industrial Electronics, 2013
- Wavelet filter-based weak signature detection method and its application on rolling element bearing prognosticsJournal of Sound and Vibration, 2006
- Model for the vibration produced by a single point defect in a rolling element bearingJournal of Sound and Vibration, 1984