This project involves two research teams, each expert in their respective areas. One team is with the University of Bordeaux, and it is skilled in distributed graph algorithms, graph theory, framework modeling, probabilistic analysis, formal methods and security, and machine learning. The other team is with MNIT Jaipur and has been working in the field of malware analysis wherein the current focus is mobile security especially in Android.
The main objective of the project is to detect privacy/sensitive data leakage through one or more (colluding) Android apps that are reflection-aware as well as equipped with anti-detection features employed for evading dynamic analysis. Such apps can not be analyzed by existing static methods alone. Android is being considered as it is open source and has largest presence in mobile OS.
This project shall aim at improving existing analysis tools through an incorporation of reflection- aware and detection-aware features in the generation of CFG (control flow graph). The main objective is to identify if an app is capable of privacy leakage on its own or in collaboration with other apps present on the device. The accuracy of such analysis methods depends on modeling of app components and their interactions as per Android semantics (which needs to be formulated from the description as no formal description by Android).