Enhancing Anomaly Detection of IoT using Knowledge-Based and Federated Deep Learning
Tabassum Simra, Bharath Konatham, Fathi Amsaad, Mohamed I. Ibrahem, Noor Zaman Jhanjhi · 2024
IoT encompasses an extensive range of sensors and physical devices that establish connections with diverse applications via networking technologies, enabling communication with the Internet and other devices. Because of the increasing number of assaults on IoT applications, it is crucial to ensure strong cybersecurity as the number of users for IoT grows and new services appear. This article examines the importance of securing IoT applications and evaluates the effectiveness of integrating knowledge distillation and federated learning techniques with Deep Learning algorithms to enhance IDS for protecting IoT applications. The utilization of a combination of knowledge distillation and federated learning showcases the potential for achieving many advantages, including enhancing model performance, expediting the learning process, and safeguarding user data privacy. These advantages have been demonstrated to surpass conventional learning approaches.