A Model of Extracting Security Situation Element Based on Federated Deep Learning for Industrial Internet

In order to solve the problems of insufficient privacy protection and limited sharing of industrial Internet security situation data,a situation element extraction model integrating federated learning and deep learning was proposed. This model integrates deep residual networks, bidirectional long short-term memory networks, and Transformer architecture,which extract features from network security situation data from multiple dimensions such as local features, temporal characteristics, and global correlations, and establish a situation element extraction model. Under the federated learning architecture,each participant performs data processing and model updates locally,transmitting model parameters through security mechanisms to reduce unnecessary data sharing and flow. The experimental results show that this method further improves the situation element extraction performance while protecting data privacy.