A survey on anomaly detection for technical systems using LSTM networks

• Published in: Computers in industry Vol. 131; p. 103498
• Main Authors: Lindemann, Benjamin, Maschler, Benjamin, Sahlab, Nada, Weyrich, Michael
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 01.10.2021
• Subjects: Anomaly detection; Artificial intelligence; Autoencoder; Context modeling; Long short-term memory; Transfer learning; Long short-term memory; Autoencoder; Transfer learning; Artificial intelligence; Anomaly detection; Context modeling
• ISSN: 0166-3615; 1872-6194
• DOI: 10.1016/j.compind.2021.103498

•Focusing on practical application of neural network-based detection algorithms.•LSTM-based approaches allow dynamic and time-variant anomaly detection.•Graph-based approaches enable unified representation of heterogeneous data.•Transfer learning addresses frequent lack of sufficiently large and diverse datasets.•Graph-based and transfer learning are promising, but still in early development. Anomalies represent deviations from the intended system operation and can lead to decreased efficiency as well as partial or complete system failure. As the causes of anomalies are often unknown due to complex system dynamics, efficient anomaly detection is necessary. Conventional detection approaches rely on statistical and time-invariant methods that fail to address the complex and dynamic nature of anomalies. With advances in artificial intelligence and increasing importance for anomaly detection and prevention in various domains, artificial neural network approaches enable the detection of more complex anomaly types while considering temporal and contextual characteristics. In this article, a survey on state-of-the-art anomaly detection using deep neural and especially long short-term memory networks is conducted. The investigated approaches are evaluated based on the application scenario, data and anomaly types as well as further metrics. To highlight the potential of upcoming anomaly detection techniques, graph-based and transfer learning approaches are also included in the survey, enabling the analysis of heterogeneous data as well as compensating for its shortage and improving the handling of dynamic processes.