CNN-Based Android Malware Detection
M. N. V. S. Ganesh, Priyanka Pednekar, Pooja Prabhuswamy, Divyashri Sreedharan Nair, Younghee Park, Hyeran Jeon · 2017
The growth in mobile devices has exponentially increased, making information easy to access but at the same time vulnerable. Malicious applications can gain access to sensitive and critical user information by exploiting unsolicited permission controls. Since high false detection rates render signature-based antivirus solutions on mobile phones ineffective, especially in malware variants, it is imperative to develop a more efficient and adaptable solution. This paper presents a deep learning-based malware detection to identify and categorize malicious applications. The proposed method investigates permission patterns based on a convolutional neural network. Our solution identifies malware with 93% accuracy on a dataset of 2500 Android applications, of which 2000 were malicious and 500 were benign.