A Survey of Android Malware Detection with Deep Neural Models

摘要: Deep Learning (DL) is a disruptive technology that has changed the landscape of cyber security research. learning models have many advantages over traditional Machine (ML) models, particularly when there large amount data available. Android malware detection or classification qualifies as big problem because fast booming number malware, obfuscation and potential protection huge values assets stored on devices. It seems natural choice to apply DL detection. However, exist challenges for researchers practitioners, such architecture, feature extraction processing, performance evaluation, even gathering adequate high quality. In this survey, we aim address by systematically reviewing latest progress in DL-based classification. We organize literature according including FCN, CNN, RNN, DBN, AE, hybrid models. The goal reveal research frontier, with focus representing code semantics also discuss emerging field provide our view future opportunities directions.