Security and privacy analysis of mobile applications data traffic using deep learning techniques

Smartphones have become an essential component of peoples’ lives where smartphone apps provide immeasurable benefits to users. However, while there are numerous benefits, activities on smartphones leak sensitive information about the users. Despite the widespread use of encryption techniques to provide confidentiality over Internet communications, mobile device users are still susceptible to privacy and security risks. In this project, we develop a testbed to simulate and collect mobile applications’ encrypted traffic data. We also develop novel windows-based feature extraction techniques and deep learning-based data classification techniques to study the privacy vulnerabilities of today’s smartphones. Notably, we exploit the probability distribution at deep learning’s output layer to filter unseen mobile application traffic to increase classification accuracy.

This research project has been ongoing since 2018, and has alignment with both the Cyber Security, Trust, Identity and Privacy and Artificial Intelligence, Machine Learning and Data Analytics research themes of the IDT.

The following publication provides further info on this research: M.H. Pathmaperuma, Y. Rahulamathavan, S. Dogan, and A.M. Kondoz, “In-App Activity Recognition from Wi-Fi Encrypted Traffic,” in: Arai K., Kapoor S., Bhatia R. (eds) Intelligent Computing, Advances in Intelligent Systems and Computing, (SAI 2020), vol 1228. Pp. pp 685-697, Springer, Cham., 2020

For further info, please contact Dr Safak Dogan.