A Survey on Adversarial Attacks for Malware Analysis
Machine learning-based malware analysis approaches are widely researched and deployed in critical infrastructures for detecting and classifying evasive and growing malware threats. However, minor perturbations or ineffectual byte insertions can easily ‘fool’ these trained ML cl...
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2025-01-01
|
| Series: | IEEE Access |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/10806701/ |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1841563317226700800 |
|---|---|
| author | Kshitiz Aryal Maanak Gupta Mahmoud Abdelsalam Pradip Kunwar Bhavani Thuraisingham |
| author_facet | Kshitiz Aryal Maanak Gupta Mahmoud Abdelsalam Pradip Kunwar Bhavani Thuraisingham |
| author_sort | Kshitiz Aryal |
| collection | DOAJ |
| description | Machine learning-based malware analysis approaches are widely researched and deployed in critical infrastructures for detecting and classifying evasive and growing malware threats. However, minor perturbations or ineffectual byte insertions can easily ‘fool’ these trained ML classifiers, making them ineffective against these crafted and smart malicious software. This survey aims to provide an encyclopedic overview of adversarial evasion attacks specifically targeting malware detection and classification systems, standing apart from previous surveys by focusing exclusively and comprehensively on this unique application domain. While significant strides have been made in adversarial research in other fields, the specific challenges of adversarial malware remain under-explored due to the intricate nature and constraints of the malware domain. Our survey addresses this gap by analyzing literature on adversarial evasion attacks published between 2013 and 2024, making it one of the first to systematically focus on malware-specific adversarial attacks in a detailed, self-contained manner. The paper will begin by introducing various machine-learning techniques used to generate adversarial malware samples, including the structural nuances of target files, which influence adversarial vulnerabilities. The work presents an in-depth threat model specific to adversarial malware evasion attacks, describing the unique attack surfaces of malware detectors and outlining adversarial goals tailored to the malware domain. We systematically analyze adversarial generation algorithms from broader domains adapted to malware evasion attacks, proposing a taxonomy of adversarial evasion attacks within malware detection based on target domains(Windows, Android and PDF). The survey highlights real-world adversarial evasion attacks on machine learning-based anti-malware engines under each taxonomical heading, demonstrating the evolution and refinement of these attack strategies over time. Our survey outlines current limitations and practical challenges in executing adversarial attacks against malware detectors in real-world environments. We identify open problems and propose future research directions for developing more practical, robust, efficient, and generalized adversarial attacks on ML-based malware classifiers. |
| format | Article |
| id | doaj-art-af186503dad84b648036f8332073875d |
| institution | Kabale University |
| issn | 2169-3536 |
| language | English |
| publishDate | 2025-01-01 |
| publisher | IEEE |
| record_format | Article |
| series | IEEE Access |
| spelling | doaj-art-af186503dad84b648036f8332073875d2025-01-03T00:01:50ZengIEEEIEEE Access2169-35362025-01-011342845910.1109/ACCESS.2024.351952410806701A Survey on Adversarial Attacks for Malware AnalysisKshitiz Aryal0https://orcid.org/0000-0001-8000-1086Maanak Gupta1https://orcid.org/0000-0001-9189-2478Mahmoud Abdelsalam2https://orcid.org/0000-0001-5627-5239Pradip Kunwar3https://orcid.org/0009-0004-2583-5925Bhavani Thuraisingham4Department of Computer Science, Tennessee Technological University, Cookeville, TN, USADepartment of Computer Science, Tennessee Technological University, Cookeville, TN, USADepartment of Computer Science, North Carolina A&T State University, Greensboro, NC, USADepartment of Computer Science, Tennessee Technological University, Cookeville, TN, USADepartment of Computer Science, The University of Texas at Dallas, Dallas, TX, USAMachine learning-based malware analysis approaches are widely researched and deployed in critical infrastructures for detecting and classifying evasive and growing malware threats. However, minor perturbations or ineffectual byte insertions can easily ‘fool’ these trained ML classifiers, making them ineffective against these crafted and smart malicious software. This survey aims to provide an encyclopedic overview of adversarial evasion attacks specifically targeting malware detection and classification systems, standing apart from previous surveys by focusing exclusively and comprehensively on this unique application domain. While significant strides have been made in adversarial research in other fields, the specific challenges of adversarial malware remain under-explored due to the intricate nature and constraints of the malware domain. Our survey addresses this gap by analyzing literature on adversarial evasion attacks published between 2013 and 2024, making it one of the first to systematically focus on malware-specific adversarial attacks in a detailed, self-contained manner. The paper will begin by introducing various machine-learning techniques used to generate adversarial malware samples, including the structural nuances of target files, which influence adversarial vulnerabilities. The work presents an in-depth threat model specific to adversarial malware evasion attacks, describing the unique attack surfaces of malware detectors and outlining adversarial goals tailored to the malware domain. We systematically analyze adversarial generation algorithms from broader domains adapted to malware evasion attacks, proposing a taxonomy of adversarial evasion attacks within malware detection based on target domains(Windows, Android and PDF). The survey highlights real-world adversarial evasion attacks on machine learning-based anti-malware engines under each taxonomical heading, demonstrating the evolution and refinement of these attack strategies over time. Our survey outlines current limitations and practical challenges in executing adversarial attacks against malware detectors in real-world environments. We identify open problems and propose future research directions for developing more practical, robust, efficient, and generalized adversarial attacks on ML-based malware classifiers.https://ieeexplore.ieee.org/document/10806701/Adversarial evasion attackadversary modelingsecurity for AIwindows PE malwareAndroid malwarePDF malware |
| spellingShingle | Kshitiz Aryal Maanak Gupta Mahmoud Abdelsalam Pradip Kunwar Bhavani Thuraisingham A Survey on Adversarial Attacks for Malware Analysis IEEE Access Adversarial evasion attack adversary modeling security for AI windows PE malware Android malware PDF malware |
| title | A Survey on Adversarial Attacks for Malware Analysis |
| title_full | A Survey on Adversarial Attacks for Malware Analysis |
| title_fullStr | A Survey on Adversarial Attacks for Malware Analysis |
| title_full_unstemmed | A Survey on Adversarial Attacks for Malware Analysis |
| title_short | A Survey on Adversarial Attacks for Malware Analysis |
| title_sort | survey on adversarial attacks for malware analysis |
| topic | Adversarial evasion attack adversary modeling security for AI windows PE malware Android malware PDF malware |
| url | https://ieeexplore.ieee.org/document/10806701/ |
| work_keys_str_mv | AT kshitizaryal asurveyonadversarialattacksformalwareanalysis AT maanakgupta asurveyonadversarialattacksformalwareanalysis AT mahmoudabdelsalam asurveyonadversarialattacksformalwareanalysis AT pradipkunwar asurveyonadversarialattacksformalwareanalysis AT bhavanithuraisingham asurveyonadversarialattacksformalwareanalysis AT kshitizaryal surveyonadversarialattacksformalwareanalysis AT maanakgupta surveyonadversarialattacksformalwareanalysis AT mahmoudabdelsalam surveyonadversarialattacksformalwareanalysis AT pradipkunwar surveyonadversarialattacksformalwareanalysis AT bhavanithuraisingham surveyonadversarialattacksformalwareanalysis |