| Fargana Abdullayeva, Gurban Farajov Cloud Cyber Attack Images Classification Using GAN and ViT+ML Algorithms |
|---|
| Abstract. The emergence of the Industry 4.0 concept and the development of modern technologies have made the detection of cyber attacks in cloud systems an important issue. In the article, a hybrid model based on the combination of machine learning algorithms with Generative Adversarial Networks (GANs) was developed to identify various attack categories targeting cloud systems. In the model, the integration of functions that enhance image quality within the GAN algorithm significantly improved classification performance by increasing the quality of cyber attack images. Here, the damage in the images is repaired, and their appearance is restored and generated to resemble the original as closely as possible. To enhance the model's robustness against various changes in input images, during the data augmentation phase, the process of rotating images and generating them in different variations was also carried out using GAN. The proposed method classified various cyber attacks on cloud systems more effectively than existing methods, achieving a classification accuracy of 0.9451. |
| Keywords: Deep learning, GANs, CatBoost, Cyber Attacks, vision transformer (ViT) |
 Download PDF |
| DOI: |
