通信学报2024，Vol.45Issue(4)：73-83,11.DOI:10.11959/j.issn.1000-436x.2024077
基于后门攻击的恶意流量逃逸方法
Escape method of malicious traffic based on backdoor attack
摘要
Abstract
Launching backdoor attacks against deep learning(DL)-based network traffic classifiers,and a method of ma-licious traffic escape was proposed based on the backdoor attack.Backdoors were embedded in classifiers by mixing poi-soned training samples with clean samples during the training process.These backdoor classifiers then identified the ma-licious traffic with an attacker-specific backdoor trigger as benign,allowing the malicious traffic to escape.Additionally,backdoor classifiers behaved normally on clean samples,ensuring the backdoor's concealment.Different backdoor trig-gers were adopted to generate various backdoor models,the effects of different malicious traffic on different backdoor models were compared,and the influence of different backdoors on the model's performance was analyzed.The effective-ness of the proposed method was verified through experiments,providing a new approach for escaping malicious traffic from classifiers.关键词
后门攻击/恶意流量逃逸/深度学习/网络流量分类Key words
backdoor attack/escape of malicious traffic/deep learning/network traffic classification分类
信息技术与安全科学引用本文复制引用
马博文,郭渊博,马骏,张琦,方晨..基于后门攻击的恶意流量逃逸方法[J].通信学报,2024,45(4):73-83,11.基金项目
国家自然科学基金资助项目(No.62276091) （No.62276091）
国家社会科学基金资助项目(No.2022-SKJJ-B-057)The National Natural Science Foundation of China(No.62276091),The National Social Science Fund of China(No.2022-SKJJ-B-057) （No.2022-SKJJ-B-057）
