UMass Database and Information Management Lab

Print this page

Private Dissemination of Social Network Data

Many phenomena are naturally modeled as networks, in which entities (represented by nodes) participate in binary relationships (represented by edges). In a social network, nodes are individuals and edges are personal contacts or relationships. In a communication network, nodes are individuals and edges are flows of information such as phone calls or Email messages. In a technological network, nodes are machines, such as computers or power stations, which are connected by a means of transmission.

By analyzing data sets representing networks, researchers have studied topics as diverse as the robustness of the Internet, the spread of HIV, the causes of financial fraud, and the processes of cellular metabolism. However, a looming barrier threatens to disrupt the growing flood of research on networks. Many of the key systems under study record personal information about participants. The pattern and content of communications, social relationships, organizational affiliations, and online behavior is often laid bare within these data sets, and many participants object to wide disclosure of this information.

The goal of this project is to understand the extent to which accurate analysis of networks can be performed while safeguarding the privacy of participants. The publications below explore threats to privacy in networked data, anonymizing transformations of network data, and output perturbation techniques for accurate network analysis.

Publications

• Accurate Estimation of the Degree Distribution of Private Networks
  Michael Hay, Chao Li, Gerome Miklau, David Jensen
  To appear, International Conference on Data Mining (ICDM) 2009
  Best Student Paper Award Winner
• Enabling Accurate Analysis of Private Network Data
  Michael Hay, Gerome Miklau, David Jensen
  Draft book chapter, Privacy-Aware Knowledge Discovery: Novel Applications and New Techniques,
  Chapman & Hall/CRC Press.
• Boosting the Accuracy of Differentially-Private Histograms Through Consistency
  Michael Hay, Vibhor Rastogi, Gerome Miklau, Dan Suciu
  To appear, Conference on Very Large Databases (VLDB) 2010
• Relationship Privacy: Output Perturbation for Queries with Joins
  Vibhor Rastogi, Michael Hay, Gerome Miklau, Dan Suciu
  Principles of Database Systems (PODS) 2009
• Resisting Structural Identification in Anonymized Social Networks
  Michael Hay, Gerome Miklau, David Jensen, Don Towsley, and Philipp Weis
  Conference on Very Large Databases (VLDB) 2008
• Anonymizing Social Networks
  Michael Hay, Gerome Miklau, David Jensen, Philipp Weis, and Siddharth Srivastava
  University of Massachusetts, Amherst, Technical Report No. 07-19
  March 2007

Project Members

• Michael Hay (graduate student)
• Chao Li (graduate student)
• Gerome Miklau (faculty)
• David Jensen (faculty)
• Don Towsley (faculty)

This project is funded in part by NSF Cybertrust grant (CNS-0627642) Preserving utility while ensuring privacy for linked data, 2006-2009.