This project will explore different deep learning techniques for Android app analyses, e.g., to detect Android malware, to identify common vulnerabilities, or to pinpoint repackaged Android apps. The project will start with exercising existing deep learning models to have a better understanding of how deep learning works in practice, so as to be able to quickly set up an appropriate environment for running deep earning-based experiments. Then, this project will investigate different representations of Android apps to prepare the inputs of deep learning models. Different inputs are expected to reach different outputs. The goal of this project is, therefore, to devise reliable inputs (i.e., representation of Android apps) for the deep learning models to analyse Android apps with high precision and recall.
Project based scholarship