2007-626 Protecting Privacy from Social Network Structure-based Inference

Summary

UCLA researchers in the Department of Computer Science have developed an invention to quantify how much a piece of personal information can be inferred based on social network structures revealed in online social networking services and methods to automatically generate recommendations to protect such personal information.

Background

Recent years have seen a huge growth in online social networking sites such as Facebook, Myspace, and Friendster. Given the huge amount of personal data and social relationships available in online social networks, protecting ones personal privacy is a growing problem. Since private information can be inferred via social relationships, it is possible to infer private information even when such information is not shared.

Innovation

The invention uses a method to infer with a high degree of accuracy personal information based on social networks.

Applications

  • For the end user: implementation in a personal privacy management and advising software tool
  • For an online social networking provider: provide privacy evaluation as part of its service to end users
  • For a product recommender: based on a users interest on the set of attributes for a given product, using the social relationships among friends to infer which of his friends will also be interested in this product.

Advantages

Currently, no social network services implement any privacy evaluation, alert, and recommendation techniques that could help end users easily evaluate and manage the level to which their private information is revealed. Neither does any personal software tool exist on the market that provides such functionality.

State of Development

The concept has been developed and results of the experiment have been published.

Patent Information:
For More Information:
Joel Kehle
Business Development Officer
joel.kehle@tdg.ucla.edu
Inventors: