I’m working on spatio-temporal data mining and privacy protection. I am particularly interested in leveraging geometric and probabilistic algorithms on efficient representation and effective mining of trajectory data, while protecting users’ privacy.
In this paper we formulate the problem of inferring locations of mobile agents, present theoretically-proven bounds on the amount of information that could be leaked in this manner, study their geometric nature, and present algorithms matching these bounds.
Boris Aronov, Alon Efrat, Ming Li, Jie Gao, Joseph S.B. Mitchell, Valentin Polishchuk, Boyang Wang, Hanyu Quan, Jiaxin Ding
In this work, we employ a ``mix-and-match'' approach to collect shorter trajectories, with which local geometric features of trajectories are preserved, while the re-identification of a user via frequent locations, co-locations and spatial-temporal points are not possible with high probability.
In this work, we employ the minhash signatures to store the trajectories with differential privacy guarantee and build a distributed data structure, MinHash hiearchy, to answer queries regarding popular paths efficiently.