Gautam Srivastava

Social Network Anonymization via Edge Addition

The growing need to address privacy concerns when social network data is released for mining purposes has recently led to considerable interest in various techniques for graph anonymization. In this paper, we study the following problem: Given a social network modeled as an {\em edge-labeled graph}

Social Network Privacy for Attribute Disclosure Attacks

G

Label-bag based Graph Anonymization via Edge Addition

, we aim to make a pre-specified subset of vertices of

SPARCS from the University of Victoria: Supporting Sustainable and Integrated Outreach Activities for Educators and Young Minds

G

Complexity of social network anonymization

k

Why Waldo befriended the dummy? k -Anonymization of social networks with pseudo-nodes

-label sequence anonymous with the minimum number of edge additions. Here, the label sequence of a vertex is the sequence of labels of edges incident to it. The contributions of this paper are two fold: We provide a framework to show hardness results for different variants of social network anonymization using a common approach. We start by showing that

k-Anonymization of Social Networks by Vertex Addition.

k

The Effects of Ant Colony Optimization on Graph Anonymization

-label sequence anonymity of arbitrary labeled graphs is hard, and use this result to prove NP-hardness results for many other recently proposed notions of graph anonymization. Secondly, we present interesting algorithms and hardness for bipartite graphs. For unlabeled bipartite graphs, we show

Graph anonymization through edge and vertex addition

k