IoT Intrusion Detection
IoT: Internet of Things
Overview
IoT intrusion detection studies how to recognize attacks such as botnets, denial-of-service, and probing in networks of connected devices and industrial systems. Approaches range from Random Forest and gradient-boosted flow classifiers to CNNs, recurrent networks, Transformers, graph models, and anomaly detection; federated and compressed models aim to work across private, resource-limited edge devices. Evaluation commonly uses datasets such as CICIoT2023, Edge-IIoTset, and ToN-IoT, but leakage and dataset-specific shortcuts can make impressive test scores overstate real-world performance. A central challenge is achieving reliable cross-device and cross-network generalization while resisting evasion or poisoning and meeting constraints on privacy, latency, and computation.
Latest papers
53 papers2 papers in the last four weeks, down 80% on the four weeks before. 0.0% of all new papers.