Adversarial Machine Learning in Cybersecurity: Attacks and Defenses
Ke Hu, Jian Feng Xu, Yong Wang, Heyao Chen, Zepeng Shen · International Journal of Management Science Research · 2025
Adversarial Machine Learning (AML) refers to the research field that involves testing and improving machine learning models by introducing adversarial samples or attack techniques. In the cybersecurity domain, AML has significant potential to help identify and defend against threats such as malware, cyber attacks, and identity fraud. However, AML also faces numerous challenges, including low efficiency in generating adversarial samples, insufficient stealth, and issues with the generality and adaptability of defense methods. There is a dynamic interplay between adversarial attacks and defenses, with attackers continually developing new techniques and defenders needing to constantly improve their defense strategies. This interaction drives the rapid development of AML technology, making it increasingly important in cybersecurity. By deeply studying the interplay between adversarial attacks and defenses, the robustness and reliability of cybersecurity systems can be effectively enhanced, laying the foundation for future AI development in cybersecurity.