联邦学习驱动的网络入侵检测研究综述OA

With the rapid growth of network traffic and the continuous evolution of attack methods,traditional intrusion detection methods have exposed obvious deficiencies in data island,privacy protection and heterogeneous environment adaptation.Federated learning provides a new idea for cross-domain collaborative modeling and privacy protection because of its characteristic of"data not being local".However,there is still a lack of systematic combing and method evolution analysis for its research in the field of intrusion detection.This paper focuses on the research of network intru-sion detection driven by federated learning.Starting from the three core dimensions of aggregation strategy,privacy pro-tection mechanism and detection technology,the paper systematically summarizes the research progress in recent years,analyzes the advantages and disadvantages of various methods and applicable scenarios,and summarizes the current chal-lenges and development trends.This paper provides research ideas for relevant researchers in this field,and also provides theoretical support for the design and deployment of actual systems.