Summary

Security is the single most important issue in IoT realization. Traditional PC/desktop security solutions, especially signature-based solutions, are not effective in IoT applications. Behavior-based solutions, such as anomaly detection, are preferable in IoT. DL/ML models are very useful tools in IoT for data analysis and also for security incident detection. In this chapter, we presented DL-based network and device behavioral data analysis, and security incident detection techniques for IoT applications in general. In the first part of this chapter, we briefly described various IoT security attacks and their potential detection techniques, including DL/ML based techniques. We considered two different levels of intrusion detection in IoT applications. The first use case is on device-level or host-level intrusion detection, and the second use case is on network-level intrusion...