- All Journals
- Al-Qadisiyah University
- Basrah University
- Kirkuk University
- Mosul University
- The Iraqi Borad for Medical Specialization
- University of Anbar
- University of Technology
- Al Mustafa University
- Al-Zahraa University for Women
- Alnoor University
- Kunooz University
- Misan University
- Ninevah University
- Tikrit University
- University of Al-Hamdaniya
- University of Kerbala
Towards Robust and Explainable AI-Based Malware Detection: A Survey of Adversarial Attacks, Defenses, and Open Challenges | ||
| Al-Noor Journal for Information Technology and Cybersecurity | ||
| Articles in Press, Accepted Manuscript, Available Online from 19 January 2026 | ||
| Document Type: Research paper | ||
| Authors | ||
| Mohammed Basim Al Abbar* 1; Najla Badie Al Dabagh2 | ||
| 1University of Mosul, College of Computer Science and Mathematics, Department of Computer Science, Iraq | ||
| 2University of Mosul, College of Computer Science and Mathematics, Department of Computer Science, Iraq, | ||
| Abstract | ||
| Machine learning (ML) and deep learning (DL) have become the hotbeds of automated feature extraction that achieve high-accuracy classification results over a wide range of platforms in modern malware detection. This makes them, however, very susceptible to adversarial manipulations since slight, benign-looking modifications can fool ML/DL classifiers while still carrying out the intended malicious activity. The survey is organized as follows: It first covers static, dynamic, and hybrid analysis techniques; feature representation methods; and major artificial intelligence models used for malware detection today; then it provides an advanced taxonomy on gradient-based attacks at the binary level and behavioral reinforcement learning-driven evasions against current advanced defenses, both proactive and reactive, discovering remaining weaknesses, robustness, scalability limitations, and locking down final public datasets benchmarks testing coverage labeling consistency and real-world representativeness. Most of the tools perform well when tested with standard datasets. However, their performance drastically reduces in the presence of an intelligent adversary or adaptive attack. Therefore, this paper highlights the need for future research to focus on developing explainable, robust, cross-domain, generalized, and adaptive defense strategy-based models for securing practical AI-based malware detections. | ||
| Keywords | ||
| Adversarial Machine Learning; AI-Based Malware Detection; Adversarial Evasion Attacks; Robust Defense Mechanisms; Malware Analysis (Static Dynamic Hybrid) | ||
|
Statistics Article View: 249 |
||