Graph Identification and Privacy in Social Networks

                           Lise Getoor
                      University of Maryland

Graph identification refers to methods that transform observational
data described as a noisy, input graph into an inferred "clean" output
graph. Examples include inferring social networks from communication
data, identifying gene regulatory networks from protein-protein
interactions, and distinguishing legitimate vs. malicious traffic from
noisy and incomplete trace route data.  On the flip-side, there is a
growing interest in anonymizing social network data, and understanding
the different types of privacy threats inherent in relational data.
In this talk, I will discuss some of the key processes involved in
identification (entity resolution, link prediction, collective
classification and group detection) and I will overview results
showing that on several well-known social media sites, we can easily
and accurately recover information that users may wish to remain
private.