A Generative Neural Network for Improving Metamorphic Malware Detection in IoT Mobile Devices

Turnbull, Leigh; Tan, Zhiyuan; Babaagba, Kehinde O.

Abstract

There has been an upsurge in malicious attacks in recent years, impacting computer systems and networks. More and more novel malware families aimed at information assets were launched daily over the past year. A particularly threatening malicious group is metamorphic malware that uses several obfuscation techniques to transform its code structure between generations. This malicious family thus poses more difficulty in its analysis and detection. In defeating metamorphic malware, several Machine Learning (ML) techniques have been employed and have been shown to outperform other conventional techniques. In this research, we examine the use of ML, a Generative Neural Network in particular, for improving metamorphic malware detection in Android Operating System (OS) (this represents the most common mobile OS) by augmenting training data. The experimental results demonstrate enhanced detection of novel metamorphic malware by augmenting training data, comprising new samples derived from Deep Convolutional Generative Adversarial Network (DCGAN) and features from metamorphic malware samples.

• Online Publication Date: Oct 26, 2023
• Publication Date: 2024
• Deposit Date: Nov 1, 2023
• Publisher: CRC Press
• Pages: 24-53
• Book Title: Internet of Things Security and Privacy: Practical and Management Perspectives
• Chapter Number: 2
• ISBN: 9781032057712, 9781032058306
• DOI: https://doi.org/10.1201/9781003199410-2
• Public URL: http://researchrepository.napier.ac.uk/Output/3224138