Network Intrusion Detection Based on Transformer Model

In this paper, we present a new technique based on the Attention Mechanism of transformer model. The proposed solution is a modified version of the transformer model which has been proposed and used in the language translation domain. We conduct experiments on a dataset containing network attacks. We have evaluated our performance results and demonstrated that the proposed model significantly has good accuracy, precision, recall, and false positive rates. The experimental results indicate that the proposed approach is efficient and effective in dealing with network attacks.