Advanced evasion attacks and mitigations on practical ML-based phishing website classifiers

doi:10.1002/int.22510

	Advanced evasion attacks and mitigations on practical ML-based phishing website classifiers
	Song, Fu1 ; Lei, Yusi 1; Chen, Sen 2; Fan, Lingling 3; Liu, Yang 4
	2021-09
发表期刊	INTERNATIONAL JOURNAL OF INTELLIGENT SYSTEMS (IF:5.0[JCR-2023],5.9[5-Year])
ISSN	0884-8173
EISSN	1098-111X
卷号	36 期号:9 页码:5210-5240
DOI	10.1002/int.22510
摘要	Machine learning (ML) based classifiers are vulnerable to evasion attacks, as shown by recent attacks. However, there is a lack of systematic study of evasion attacks on ML-based anti-phishing detection. In this study, we show that evasion attacks are not only effective on practical ML-based classifiers, but can also be efficiently launched without destructing the functionalities and appearance. For this purpose, we propose three mutation-based attacks, differing in the knowledge of the target classifier, addressing a key technical challenge: automatically crafting an adversarial sample from a known phishing website in a way that can mislead classifiers. To launch attacks in the white- and gray-box scenarios, we also propose a sample-based collision attack to gain the knowledge of the target classifier. We demonstrate the efficacy of our evasion attacks on the state-of-the-art, Google's phishing page filter, achieved 100% attack success rate in less than one second per website. Moreover, the transferability attack on BitDefender's industrial phishing page classifier, TrafficLight, achieved up to 81.25% attack success rate. We further propose a similarity-based method to mitigate such evasion attacks, Pelican, which compares the similarity of an unknown website with recently detected phishing websites. We demonstrate that Pelican can effectively detect evasion attacks, hence could be integrated into ML-based classifiers. We also highlight two strategies of classification rule selection to enhance the robustness of classifiers. Our findings contribute to design more robust phishing website classifiers in practice.
关键词	adversarial attacks adversarial sample detection machine learning mutation phishing website
URL	查看原文
收录类别	SCIE ; EI
语种	英语
WOS研究方向	Computer Science
WOS类目	Computer Science, Artificial Intelligence
WOS记录号	WOS:000661142100001
出版者	WILEY
原始文献类型	Article; Early Access
引用统计	正在获取...
文献类型	期刊论文
条目标识符	https://kms.shanghaitech.edu.cn/handle/2MSLDSTB/127570
专题	信息科学与技术学院_PI研究组_宋富组
通讯作者	Song, Fu
作者单位	1.ShanghaiTech Univ, Sch Informat Sci & Technol, 393 Huaxia Middle Rd, Shanghai 201210, Pudong, Peoples R China; 2.Tianjin Univ, Sch Cybersecur, Coll Intelligence & Comp, Tianjin, Peoples R China; 3.Nankai Univ, Coll Cyber Sci, Tianjin, Peoples R China; 4.Nanyang Technol Univ, Sch Comp Sci & Engn, Singapore, Singapore
第一作者单位	信息科学与技术学院
通讯作者单位	信息科学与技术学院
第一作者的第一单位	信息科学与技术学院
推荐引用方式 GB/T 7714	Song, Fu,Lei, Yusi,Chen, Sen,et al. Advanced evasion attacks and mitigations on practical ML-based phishing website classifiers[J]. INTERNATIONAL JOURNAL OF INTELLIGENT SYSTEMS,2021,36(9):5210-5240.
APA	Song, Fu,Lei, Yusi,Chen, Sen,Fan, Lingling,&Liu, Yang.(2021).Advanced evasion attacks and mitigations on practical ML-based phishing website classifiers.INTERNATIONAL JOURNAL OF INTELLIGENT SYSTEMS,36(9),5210-5240.
MLA	Song, Fu,et al."Advanced evasion attacks and mitigations on practical ML-based phishing website classifiers".INTERNATIONAL JOURNAL OF INTELLIGENT SYSTEMS 36.9(2021):5210-5240.