APPLYING DEEP LEARNING TECHNIQUES FOR ANDROID MALWARE DETECTION

This article explores the use of deep learning for malware identification in the Android operating system. Similar studies are considered and, based on their drawbacks, a self-designed approach is proposed for representing an Android application for a convolutional neural network, which consists in constructing an RGB image, the pixels of which are formed from a sequence of pairs of API calls and protection levels. The results of the experimental evaluation of the proposed approach, which are presented in this paper, demonstrate its high efficiency for solving the problem of identifying malicious Android applications.