Title: Privacy in geo-social networks

Contact person: Dr. Jun Pang, Yang Zhang

Short description:
Social networks, such as Facebook, Twitter and Linkedln, are playing a
great role in our daily life. They provide a novel platform to
participants to make new friends, maintain close connection to old
friends and exchange status. In recent years, motivated by the
increased popularity of GPS-enabled mobile devices, social networks
have evolved into geo-social networks which support location-based
services. People can tag messages and photos with their geographical
locations, find nearby friends and post check-in of some places to
share their comments. (Besides those location-sharing social networks
such as Foursquare and Google Latitude, almost all existing large
social networks, such as Facebook and RenRen, have upgraded to support
location-based applications.)

As we know, privacy and information sharing are two sides of a
coin. Social networks cannot be exceptional either, especially with
shared locations which can reveal private information. For instance,
when a daily trace of locations are collected, with the help of other
public information, such as maps and yellow pages, a user's daily
activities could be reconstructed even though no messages are posted
explicitly about them. Moreover, people tend to use real names
registered as identities, which makes the privacy protection in social
networks is not only necessary but also much more challenging.

This project is designed to model and analyze the privacy risks in
some specific applications of geo-social networks, find new
protections of users' location privacy and prevent information leakage
from published datasets.  So far, there are some methodologies being
widely utilized in the literature, involving graph-based algorithms,
clustering-based analysis and mathematical theories such as
probability theory and topological analysis. It is also interesting to
integrate the existing solutions in other domains such as voting and
trust management. The topic of the project is open and will be fixed
according to the background of the selected students.