Use of GAN Models and CNN Heatmaps in Malware
摘要
Malware has become an increasingly critical issue in the field of cybersecurity, as the latest malware techniques are designed to evade most of the malware detectors nowadays, resulting in the traditional security techniques becoming obsolete. Some of the current techniques are developed through the use of neural networks. These include malware generation using Generative Adversarial Network (GAN) models, and the use of Convolutional Neural Network (CNN) models. This survey will present the current state of both CNN and GAN model frameworks, and the potential advances that the combination of both models can bring to the current context of the malware field.