Literature Review on Improving IoT Security Using Machine and Deep Learning Technique

Abstract

This literature review paper discusses the concept of the Internet of Things (IoT) and its implications on various domains, highlighting the challenges and security concerns associated with its expansive scope. This review work emphasizes the need for comprehensive security solutions to address the complexities of IoT infrastructure, particularly in the context of emerging threats. Thispaper also underscores the importance of integrating security, energy efficiency, software applications, and data analytics in IoT systems. It outlines the evolving landscape of IoT security, including the vulnerabilities and potential consequences of inadequate security measures. Additionally, authors address the intersection of security and privacy concerns within Deep Learning (DL) and Machine Learning (ML), discussing various strategies such as homomorphic encryption, differential privacy, trusted execution, and secure multiparty computing. It acknowledges the computational demands of these approaches and the ongoing search for globally harmonized solutions. Finally, authors concludes by highlighting the challenges and strategies in countering adversarial attacks in DL and ML, emphasizing the effectiveness of adversarial training and the multifaceted nature of defense mechanisms.