Geo-Indistinguishability: Differential Privacy for Location-Based Systems
University:Ecole Polytechnique of Paris
Room :Αίθουσα Συνεδριάσεων Α56 (1ος όροφος)
In this talk I will present a line of work aimed at protecting the privacy of a
user when dealing with location-based services. The starting point of our
approach is the principle of geo-indistinguishability, a formal notion of
privacy, inspired by the concept of Differential Privacy, that protects the
user's exact location, while allowing approximate information -- typically
needed to obtain a certain desired service -- to be released.
I will focus on a few results in this area, while giving a brief overview of
some others. These include: a generic efficient mechanism to sanitize locations
with reasonable utility; a custom-built mechanism for a limited set of locations
but providing optimal utility; a method to limit the privacy degradation in case
of repeated use of the mechanism; a technique to construct elastic privacy
metrics taking into account the semantics of each location; and methods for
substantially improving the utility of location obfuscation, while having
practical applicability as a central constraint.
Finally, I will present Location Guard, a web browser extension that provides
location privacy when using the HTML5 geolocation API. The extension has reached
considerable popularity since its release, with more than 60k weekly active users.