Date of Award
Campus Access Dissertation
Doctor of Philosophy (PhD)
The past decade has witnessed tremendous growth in the mobile computing arena. Mobile users are able to participate in exciting social media applications, receiving information tailored to their location. However, there are growing privacy concerns that hinder further development of location-based applications as raw location data can disclose sensitive information about individuals. This thesis presents solutions for privacy-preserving location-processing algorithms. Both online and offline scenarios are considered: Two main directions are in the context of differential privacy, the de facto standard for statistical privacy-preserving data release: privately analyzing crowdsourced data collected by the users and extracting significant patterns from user trajectories. A third direction considers the case of online, non-statistical queries and our focus is to optimize a private information retrieval protocol by using GPU programming. The proposed solutions achieve strong protection, are computationally efficient and maintain high data accuracy, sometimes by more than an order of magnitude compared to competitive approaches.
Maruseac, Mihai, "Effective Online and Offline Privacy-Preserving Processing of Location Data" (2016). Graduate Doctoral Dissertations. 296.