ABSTRACT
Internet of Things (IoT), an interconnection of devices over the Internet, is one of the biggest expansions of networks. There has been a rapid growth of the IoT sector in both academia and industry, which has helped to discover and flag numerous concerns pertaining to privacy and security. Since IoT systems function by tracking habits, behaviors, geolocations, and other personal and potentially confidential data over an extended period of time, they must ensure strong data privacy and resolve security vulnerabilities such as malware, jamming, and spoofing. Machine learning has already been extensively used in a variety of fields, including fraud detection, speech recognition, and bioinformatics. It is primarily used for optimizing the performance of computer models and computational methods by detecting patterns in the data. Machine learning techniques are employed to improve various aspects of IoT security (access control, authentication, secure offloading, and malware detection). While there exist drawbacks such as computational overhead and partial state observations, the advantages of machine learning implementation in IoT security schemes far outweigh them. The secure application of IoT devices has displayed promising advancements in healthcare, smart grids, wearables, smart home, and alike.
