Deep Learning Technology: Enabling Safe Communication via the Internet of Things

Ramiz Salama ^1* Hitesh Mohapatra ^2* Tuğşat Tülbentçi ¹ Fadi Al-Turjman ¹

• ¹ Near East University, Nicosia, Cyprus
• ² KIIT University, Bhubaneswar, India

The Internet of Things (IoT) is a new type of technology that connects billions of devices. Although the devices offer many advantages, their diversified architecture and wide connectivity make them vulnerable to various cyberattacks, which could result in data breaches and financial loss. It is vital to prevent such assaults on the IoT ecosystem. This paper provides an softwaredefined networks (SDN)-enabled solution for vulnerability discovery in Internet of Things systems based on deep learning. The most recent Cuda-deep neural network, Cuda-bidirectional long short-term memory (Cu-BLSTM), and Cuda-gated recurrent unit (Cu-DNNGRU) classifiers are used for successful threat detection. A 10-fold cross-validation was used to demonstrate the findings' impartiality. We use the most recent publicly available CICIDS2021 data set to train our hybrid model. With a 99.96% recall rate and a 99.87% accuracy, the recommended procedure is effective. In addition, we compare the proposed hybrid model and existing benchmark classifiers to the long short-term memory (LSTM) of the Cuda-Deep Neural Network and Cuda-Gated Recurrent Unit (Cu-GRULSTM and Cu-DNNLSTM, respectively). Our proposed technique performs admirably in terms of evaluation criteria such as F1-score, speed efficiency, accuracy, and precision, among others.