Unsupervised Representation Learning for Anomaly Detection in Industrial IoT Systems
Download PDFAbstract
The Industrial Internet of Things (IIoT) has enabled unprecedented levels of monitoring and control in modern industrial systems. However, the massive volume of high-dimensional sensor data generated by these systems presents significant challenges for traditional anomaly detection methods. This chapter explores the application of unsupervised representation learning as a powerful paradigm for identifying anomalous behavior in IIoT environments without the need for labeled data. We introduce a methodology centered on the Variational Autoencoder (VAE), a deep generative model capable of learning a compressed, low-dimensional representation of normal system behavior. Anomalies are then detected as data points that the trained model fails to reconstruct accurately, as indicated by a high reconstruction error. This chapter details the entire workflow, from synthetic IIoT data generation and model implementation to results evaluation. Through a simulated case study, we demonstrate the effectiveness of the VAE-based approach, achieving high recall and a strong ROC-AUC score, proving its capability to identify novel and complex anomalies in industrial settings. The discussion highlights the model’s performance, the significance of its learned latent representations, and the practical implications for predictive maintenance and system reliability.
References
- K Schwab. “The Fourth Industrial Revolution, Crown Business, New York”. In: The smart-up ecosystem: Turning Open Innovation into smart business (2017).
- Ane Bl´azquez-Garcia et al. “A review on outlier/anomaly detection in time series data”. In: ACM computing surveys (CSUR) 54.3 (2021), pp. 1–33.
- Varun Chandola, Arindam Banerjee, and Vipin Kumar. “Anomaly detection: A survey”. In: ACM computing surveys (CSUR) 41.3 (2009), pp. 1–58.
- Diederik P Kingma and Max Welling. “Auto-encoding variational bayes”. In: arXiv preprint arXiv:1312.6114 (2013).
- Douglas C Montgomery. Introduction to statistical quality control. John wiley & sons, 2020.
- Martin Ester et al. “A density-based algorithm for discovering clusters in large spatial databases with noise”. In: kdd. Vol. 96. 34. 1996, pp. 226–231.
- Bernhard Sch¨olkopf et al. “Estimating the support of a high-dimensional distribution”. In: Neural computation 13.7 (2001), pp. 1443–1471.
- Mayu Sakurada and Takehisa Yairi. “Anomaly detection using autoencoders with nonlinear dimensionality reduction”. In: Proceedings of the MLSDA 2014 2nd workshop on machine learning for sensory data analysis. 2014, pp. 4–11.
- Jinwon An and Sungzoon Cho. “Variational autoencoder based anomaly detection using reconstruction probability”. In: Special lecture on IE 2.1 (2015), pp. 1–18.
- Pankaj Malhotra et al. “Long short term memory networks for anomaly detection in time series”. In: Proceedings. Vol. 89. 9. 2015, p. 94.
- Poornaiah Billa et al. “Detecting Faces in Noisy Images using Hit-Miss Transform (HMT)”. In: International Journal of Recent Technology and Engineering (IJRTE) 8.4 (2019), pp. 10335–10338.
- Ailin Deng and Bryan Hooi. “Graph neural network-based anomaly detection in multivariate time series”. In: Proceedings of the AAAI conference on artificial intelligence. Vol. 35. 5. 2021, pp. 4027–4035.
- Milad Memarzadeh, Bryan Matthews, and Ilya Avrekh. “Unsupervised anomaly detection in flight data using convolutional variational auto-encoder”. In: Aerospace 7.8 (2020), p. 115.
