Deep Learning for Anomaly Detection in Industrial Networks
Ayush Tripathi, Prashant Upadhyay, Pawan Kumar Goel · Advances in computational intelligence and robotics book series · 2025
As industrial networks have become increasingly complex, they have become the target of choice for cyber attacks, and thus there is a need for sophisticated anomaly detection mechanisms. This chapter delves into deep learning-based solutions to detect and mitigate cyber attacks in Industrial IoT (IIoT) and Industrial Control Systems (ICS). Utilizing methods such as autoencoders, recurrent neural networks (RNNs), and convolutional neural networks (CNNs), deep learning systems are able to identify anomalies from normal network operations in real time. The chapter covers supervised and unsupervised learning methods, feature engineering's role, the challenges posed by dataset availability, adversarial attacks, and explainability in industrial anomaly detection. Industrial applications and case studies illustrate how deep learning improves industrial cybersecurity through adaptive, scalable, and smart threat detection