Edge devices, integral to the Internet of Things (IoT) and modern smart applications, have increasingly relied on artificial intelligence (AI) for efficient operation and decision-making. However, these AI-based systems are vulnerable to adversarial attacks, which can compromise their functionality and security.
This project aims to investigate the vulnerability of machine learning models deployed on edge devices and develop robust adversarial learning techniques to defend against such attacks. The focus will be on identifying potential weaknesses in current edge computing frameworks and leveraging adversarial learning, edge computing and game theory to create novel methods to enhance their resilience.
This includes developing algorithms capable of detecting and mitigating adversarial attacks, ensuring that edge devices can maintain secure and reliable operations even under hostile conditions. Our research will cover various applications of edge devices, including real-time data processing, autonomous decision-making, and network intrusion detection.
By addressing these areas, we aim to create a framework that not only secures individual edge devices but also enhances the overall security of interconnected IoT systems. The outcomes of this project will provide critical insights and methodologies for deploying secure, AI-driven edge solutions, thereby fostering safer and more reliable IoT ecosystems.
