Internet of Things (IoT) is the fast-developing field in computing technology and plays an important role in various real-life smart applications. IoT technology refers to the collective networks of connected devices that simplify the communication between devices and clouds, as well as between the devices themselves. IoT combines billions of smart devices to access, connect and store the information in the network of any user from anywhere. Major challenges faced by IoT are security and privacy. Advanced methods to enhance the security measures of IoT systems are machine learning(ML) and deep learning(DL). Advantages of deep learning(DL) over machine learning (ML) in IoT are high performance in large datasets, automatic extraction of complex representations, and produce enhanced state-of-the-art applications.
Deep learning(DL) based IOT allows the deep linking of the IOT environment, a unified protocol to control the IoT-based devices. Deep neural networks in DL-based IoT are categorized as supervised learning (discriminative), unsupervised learning (generative), and hybrid DL. Convolutional neural networks (CNNs), recurrent neural networks (RNNs), deep autoencoders (AEs), deep belief networks (DBN), restricted Boltzmann machines (RBMs), generative adversarial networks (GANs), and the ensemble of DL networks (EDLNs) are DL algorithms utilized in DL based IoT systems.
Application areas of IoT are smart healthcare, smart transportation, smart governance, smart agriculture, smart grid, smart homes, and smart supply chain. Recent developments and research areas of DL-based IoT systems are DL-based IoT for adversarial attacks, malicious detection, anomaly detection and intrusion using DL in IoT environment, Cryptanalysis in IoT using deep learning, and lightweight DL model of onboard security systems for IoT devices.