Enhancing the privacy and security of IoT intrusion detection systems through federated learning

• Grant number: 23/07184-1
• Support Opportunities: Scholarships in Brazil - Doctorate
• Start date: October 01, 2023
• Status: Discontinued
• Field of knowledge: Engineering - Electrical Engineering - Telecommunications
• Principal Investigator: João Henrique Kleinschmidt
• Grantee: Ogobuchi Daniel Okey
• Host Institution: Centro de Engenharia, Modelagem e Ciências Sociais Aplicadas (CECS). Universidade Federal do ABC (UFABC). Ministério da Educação (Brasil). Santo André , SP, Brazil
• Associated scholarship(s): 24/12903-0 - Federated device fingerprinting and identification for enhanced intrusion detection in evolving IoT systems, BE.EP.DR

Abstract

Information transmission in the Internet of Things (IoT) era, which has highly infiltrated virtuallyall facets of human endeavours, has become an issue of grave importance. The privacy and security of IoT systems have now become crucial areas of research, considering the increasing number of vulnerabilities that cybercriminals exploit to inflict harm on these systems. Machine learning (ML) and deep learning (DL) use centralized learning approaches in developing models for IoT security. Despite the performance of the centralized learning process, major concerns such as privacy, data ownership, and high computational costs still exist. Federated learning (FL) is a method introduced recently to address the weaknesses observed in centralized learning, where sensitive IoT data for model development is stored in a particular system. FL offers the advantages of maintaining the privacy of client data, providing more secure data transmission, and reducing network bandwidth,among others. Notwithstanding, several issues, including communication efficiency, data privacy,explainability, real-time response, and scalability of FL, still pose challenges to deploying the FLbased intrusion detection system (IDS) models. The objective of this proposal is to develop an IDS solution that is optimized for both privacy and network communication efficiency in the IoT context. Our proposed solution involves enhancing the algorithm and model through compression and quantization techniques, which would address the communication overhead cost and improve interactions between end devices. In addition, we prioritize privacy-centric measures, such as differential privacy and secure multiparty computation in the algorithm design. To ensure realworld applicability, we will explore explainable artificial intelligence (XAI) algorithms in the FL context, highlighting their real-time functionality. Ultimately, our goal is to offer a comprehensive and robust solution to the privacy and communication challenges that arise in FL for IoT networks. (AU)