Fvading Deep Learning -Based Malware Detectors via Obfuscation: A Deep Reinforcement Learning Approach

被引：0
作者：
Etter, Brian [1 ]
Hu, James Lee [1 ]
Ebrahimi, Mohammadreza [2 ]
Li, Weifeng [3 ]
Li, Xin [4 ]
Chen, Hsinchun [1 ]
机构：
[1] Univ Arizona, Dept Management Informat Syst, Tucson, AZ 85721 USA
[2] Univ S Florida, Sch Informat Syst & Management, Tampa, FL USA
[3] Univ Georgia, Dept Management Informat Syst, Athens, GA USA
[4] Univ Arizona, Dept Comp Sci, Tucson, AZ USA
来源：
23RD IEEE INTERNATIONAL CONFERENCE ON DATA MINING, ICDM 2023 | 2023年
基金：
美国国家科学基金会;
关键词：
Adversarial Robustness; Reinforcement Learning; Adversarial Malware Variants; Adversarial Malware Generation; Obfuscation;
D O I：
10.1109/ICDM58522.2023.00019
中图分类号：
TP18 [人工智能理论];
学科分类号：
081104 ; 0812 ; 0835 ; 1405 ;
摘要：
Adversarial Malware Generation (AMG), the generation of adversarial malware variants to strengthen Deep Learning (DL) -based malware detectors has emerged as a crucial tool in the development of proactive cyberdefense. However, the majority of extant works offer subtle perturbations or additions to executable tiles and do not explore full-file obfuscation. In this study, we show that an open-source encryption tool coupled with a Reinforcement Learning (RL) framework can successfully obfuscate malware to evade state-of-the-art malware detection engines and outperform techniques that use advanced modification methods. Our results show that the proposed method improves the evasion rate from 27%-49% compared to widely used state-of-the-art reinforcement learning-based methods.
引用
收藏
页码：101 / 109
页数：9
相关论文
共 36 条
[1]  
Aghakhani H., When Malware is Packin' Heat
[2]  
Limits of Machine Learning Classifiers Based on Static Analysis Features
[3]   A State-of-the-Art Survey on Deep Learning Theory and Architectures [J].
Alom, Md Zahangir ;
Taha, Tarek M. ;
Yakopcic, Chris ;
Westberg, Stefan ;
Sidike, Paheding ;
Nasrin, Mst Shamima ;
Hasan, Mahmudul ;
Van Essen, Brian C. ;
Awwal, Abdul A. S. ;
Asari, Vijayan K. .
ELECTRONICS, 2019, 8 (03)
[4]  
Anderson H.S., arXiv
[5]  
Badhwar R., Polymorphic and metamorphic malware. the ciso's next frontier
[6]  
Baltazar P., How and where do hackers hide malware? malwarefox
[7]  
Bergenholtz E., Information and Communications Security, P36
[8]  
Carlini N, 2019, Arxiv, DOI arXiv:1902.06705
[9]  
Castro RL, 2019, 5TH INTERNATIONAL CONFERENCE ON INFORMATION MANAGEMENT (ICIM 2019), P20, DOI [10.1109/INFOMAN.2019.8714698, 10.1109/infoman.2019.8714698]
[10]   Adversarial Examples for CNN-Based Malware Detectors [J].
Chen, Bingcai ;
Ren, Zhongru ;
Yu, Chao ;
Hussain, Iftikhar ;
Liu, Jintao .
IEEE ACCESS, 2019, 7 :54360-54371
1234