Independent Study Presentation

Independent Study Presentation


Title:  Differentially Private Trajectories

Presenter:  Jared Murphy

Advisor:  Dr. Qijun Gu

Date/Time:  Monday, May 3rd @10:00 a.m. (CDT)

Zoom link:

Meeting ID: 999 8933 1021



In this study we explore differential privacy applied to trajectory datasets. The main idea of differential privacy is that if small arbitrary changes are made to the dataset, then an adversary cannot infer much about any single individual within the dataset, therefore providing privacy to all individuals in the dataset. This subject is extremely important in today’s society because of the massive amount of information being made public by various entities, e.g., governments releasing demographic information, social media releasing user statistics, cities releasing traffic data to improve public transport, etc. In this study, we examine an existing method to differentially privatize mobility trace data and find that is ineffective at successfully privatizing datasets of different sizes. We propose to implement an improved algorithm by creating a different tree structure. We develop a few key components towards the new algorithm, including analysis of the mobility trace data and a preliminary implementation of the algorithm.




Deadline: May 4, 2021, midnight