AMalNet: A deep learning framework based on graph convolutional networks for malware detection

Abstract

The increasing popularity of Android apps attracted widespread attention from malware authors. Traditional malware detection systems suffer from some shortcomings; computationally expensive, insufficient performance or not robust enough. To address this challenge, we (1) build a novel and highly reliable deep learning framework, named AMalNet, to learn multiple embedding representations for Android malware detection and family attribution, (2) introduce a version of Graph Convolutional Networks (GCNs) for modeling high-level graphical semantics, which automatically identifies and learns the semantic and sequential patterns, (3) use an Independently Recurrent Neural Network (IndRNN) to decode the deep semantic information, making full use of remote dependent information between nodes to independently extract features. The experimental results on multiple benchmark datasets indicated that the AMalNet framework outperforms other state-of-the-art techniques significantly.