Leveraging Machine Learning and Deep Learning in <scp>IoT</scp> Security: A Review

ABSTRACT The IoT is rapidly expanding due to constant advancements and connectivity innovations, which allow more objects to interact with one another seamlessly, revolutionizing businesses and daily life. This technology improves our lives by making them more convenient, safe, and productive. However, despite their limited resources, IoT devices seamlessly integrate into vast networks of diverse devices. Despite producing large volumes of data, these devices have major constraints, including limited memory, power, and computing capacity. These limitations make IoT devices particularly vulnerable to a variety of security attacks, making them targets for cyberattacks. As the nature of these threats changes, it becomes critical to improve IoT device security to protect against new vulnerabilities. This study presents five research inquiries related to the security of IoT, with a specific emphasis on the use of deep learning (DL) and machine learning (ML) techniques. We expect significant research patterns from a comprehensive analysis of current IoT security literature to influence future studies in this field. The increase in global IoT attacks highlights the need to develop models that use sophisticated ML/DL approaches. These algorithms' ability to identify IoT attacks in real or near‐real time depends on their accuracy and efficiency. In addition, this literature review reviews the current privacy and security issues, evaluates the limitations of existing security solutions, and explores the relevant datasets related to IoT security. We believe that this literature review will provide a comprehensive roadmap for researchers and industry professionals focusing on enhancing IoT security through the application of ML and DL technologies.