Adapting federated learning-based AI models to dynamic cyberthreats in pervasive IoT environments
S. Tamizharasi, P. E. Rubini, S. Saravana Kumar, Daniel Arockiam · 2024
In recent times, with the greater technological advancements, there has been a growing wave of cyberattacks potentially impacting various business networks and organizations. Even a simple cyberattack may significantly affect the business value of various organizations. Furthermore, the dynamic nature of the cyberthreats creates the requirement for continuous monitoring, adaptation, and vigilance. On the other hand, digital connectivity at present has become so ubiquitous, thereby the significance has become overstated. Some of the most common cyberthreats involve phishing attacks, ransomware, malware, and advanced persistent cyberthreats. In this context, the traditional cybersecurity protocols are effective to some extent but it is often complicated to maintain a secure environment with the emerging dynamic threats imposed by malicious actors. Hence, due to the increasing risk of dynamic cyberthreats, it has become crucial to seek innovative solutions to enhance security measures. Here comes the role of artificial intelligence in meeting the pressing needs of cyberthreats. Artificial intelligence, specifically the machine learning models has emerged as a cutting-edge solution for cybersecurity. The effective implementation of AI models in cybersecurity enables the appropriate analysis of large amounts of data, using which we can identify patterns and make more informed decisions in real time. Finding various AI models can assist in the detection and prevention of numerous cyberthreats effectively. Some of the significant benefits of employing AI models for cyberthreats detection include advanced threat detection, anomaly detection and behavior analysis, dynamic threat prevention, adaptive authentication, and many more. Furthermore, the design and implementation of various AI models for threat detection offer robust solutions and act as a resilient protocol against the ever-growing landscape of cyberthreats. This chapter explores various AI models to predict emerging dynamic cyberthreats and to ensure continuity and stability in their operations. This chapter also assesses the performance of every AI-based model with real-time datasets and explores the more efficient and suitable model to deal with cyberthreats.