Deep Learning in IoT Intrusion Detection

• Published in: Journal of network and systems management Vol. 30; no. 1
• Main Authors: Tsimenidis, Stefanos, Lagkas, Thomas, Rantos, Konstantinos
• Format: Journal Article
• Language: English
• Published: New York Springer US 01.01.2022; Springer Nature B.V
• Subjects: Communications Engineering; Computer Communication Networks; Computer Science; Computer Systems Organization and Communication Networks; Cybersecurity; Cybersecurity Management in the Era of AI; Deep learning; Economic impact; Electronic devices; Human influences; Impact analysis; Information Systems and Communication Service; Internet of Things; Intrusion detection systems; Networks; Operations Research/Decision Theory; Smart sensors; Deep learning; IDS; Cybersecurity; Internet of Things; Deep Neural Networks; Intrusion detection
• ISSN: 1064-7570; 1573-7705
• DOI: 10.1007/s10922-021-09621-9

The Internet of Things (IoT) is the new paradigm of our times, where smart devices and sensors from across the globe are interconnected in a global grid, and distributed applications and services impact every area of human activity. With its huge economic impact and its pervasive influence over our lives, IoT is an attractive target for criminals, and cybersecurity becomes a top priority for the IoT ecosystem. Although cybersecurity has been the subject of research for decades, the large-scale IoT architecture and the emergence of novel threats render old strategies largely inefficient. Deep learning may provide cutting edge solutions for IoT intrusion detection, with its data-driven, anomaly-based approach and ability to detect emerging, unknown attacks. This survey offers a detailed review of deep learning models that have been proposed for IoT intrusion detection. Solutions have been classified by model in a comprehensive, structured analysis of how deep learning has been applied for IoT cybersecurity and their unique contributions to the development of effective IoT intrusion detection solutions.