Tabular Federated Learning to detect cyber faults in smart buildings

By incorporating artificial intelligence into applications, everyone in the current circumstance has become intelligent, thereby reducing the burden of incessant interruption or human control. In today’s environment, most of the hardware infrastructure in buildings is connected to the internet, transforming the building infrastructure into a smart building infrastructure. In a similar vein, it should be noted that intelligent building infrastructure is susceptible to cyber-induced defects, and the issue of data privacy is a significant worry within the realm of collaborative learning. The primary objective of this study is to devise and implement an Attentive Interpretable Tabular-based Federated Learning methodology for safeguarding data privacy while detecting cyber-induced problems in the infrastructure of intelligent buildings. In this paper, Federated Learning ecosystem–based deep learning models are used to find and describe cyber-induced faults and vulnerabilities. The suggested system will be evaluated using various measures, including accuracy, precision, recall, and losses. In addition, the built ecosystem is examined using various data distributions to determine whether the outcomes are stable.