Transformers Architecture Oriented Intrusion Detection Systems: A Systematic Review

In recent years, the field of intrusion detection systems (IDS) has witnessed a paradigm shift with the emergence of deep learning techniques. Among these, Transformers has emerged as a promising architecture that exhibits remarkable capabilities in various natural language processing and computer vision tasks. In this paper, a comprehensive review was conducted to systematically evaluate the publications on Intrusion Detection Systems with Transformers Architecture. Transformers are becoming a popular choice for IDS because they can analyze sequences of data like network traffic, and identify intricate patterns that might indicate an attack. 697 papers were found using the systematic review (mapping) method to evaluate publications related to this article, which has been found to be increasingly used in Intrusion Detection Systems. The aim of our study is to identify how the Transformers architecture is used in combination with other algorithms and optimization methods, what criteria are taken into account in these choices, and the most searched topics in Transformers architecture and intrusion detection. Scientific papers published in the last 10 years were searched in Wiley, Science Direct, IEEE Explorer, ACM Digital Library and Scopus databases.