Abstract

Delving into the development of an AI-powered Network Intrusion Detection System (NIDS) designed for near-real-time threat detection in IoT networks. It explores the challenges of using outdated datasets for training intrusion detection models and proposes the use of a modern, in-house dataset reflecting current network traffic patterns and vulnerabilities. The chapter details the architecture and implementation of a Deep Learning (DL) model based on Convolutional Neural Networks (CNNs), trained on this dataset to accurately classify network traffic as benign or malicious. The proposed NIDS is evaluated in a realistic network environment, demonstrating its effectiveness in detecting intrusions in near-real time.