Towards Optimizing Malware Detection: An Approach Based on Generative Adversarial Networks and Transformers

Nowadays, cybercriminals are carrying out many forms of cyberattacks. Malware attacks, in particular, have emerged as one of the most challenging concerns in the cybersecurity area, as well as a key weapon used by cybercriminals. Malware is a term used to describe harmful software. Malware can be used to modify or destroy data on target computers, steal private information, control systems to attack other devices, host and disseminate illicit material, and disrupt vital infrastructures. As a result, many tools and approaches for detecting and mitigating malware attacks have been developed. Despite the improvement and rapid expansion of malware defense techniques, cybercriminals are able to develop more sophisticated and advanced malware that can defeat state-of-the-art security and anti-malware solutions. This paper proposes a novel approach based on generative adversarial networks and transformers to improve malware detection performance. By using generative adversarial transformers, the proposed approach aims to increase the malware data size and solve the data imbalance distribution issue. Promising experimental results showed an improved accuracy of malware detection of 3% using several pre-trained models when solving the problem of unbalanced data.