Addressing Security and Computation Challenges in IoT Using Machine Learning

The Internet of things (IoT) is widely used to implement different applications like smart home, smart health care, smart city, and smart farming system. The development of large smart devices/sensors enables smart technologies making it possible to implement the smart application in real time. The IoT system has security challenges like authentication, data privacy, access control, and intrusion detection system. Similarly, the computation of the sensed information from the environment is a challenging task. The computation must perform using distributed or decentralized architecture to overcome the centralized system difficulty. In a distributed/decentralized system when multiple nodes participate in a computational process, there is the risk of mutual consensus problem, malicious node detection, or data modification attacks. In this paper, the authors have identified machine learning as a solution to address some of the existing security and computational challenges. The paper also explained the implementation platform available for the integration of IoT with machine learning.