The impact of artificial intelligence on enhancing cyber security

Abstract

Due to the increasing number of connectivity devices, the IIoT has raised lot of security problems. The IoT networks need to be secure and efficient for the sustenance of such networks. We provide an amalgamation of FL and optimization methods for anomaly detection in IIoT. Integrating CNNs into the framework could boost the model accuracy and meanwhile reduce communication cost. Differential privacy mechanisms are adopted to provide data security and protect user privacy. The framework supports high accuracy, superior precision and recall and manages the training efficiently. The strongness and scalability of the proposed method provide a better proposition for addressing security issues in IIoT compared to central machine learning based approaches.