A Framework for the Analysis of Attacks Against Social Tagging Systems

J.J. Sandvig, Runa Bhaumik, Maryam Ramezani, Robin Burke, Bamshad Mobasher

Social tagging systems provide an open platform for users to share and annotate their resources such as photos and URLs. Due to their open nature, however, these systems present a security problem. Malicious users may try to distort the system's behavior by inserting erroneous or misleading annotations, thus altering the way in which information is presented to legitimate users. This paper addresses the problem of modeling attacks against social tagging systems and evaluating their impact on the systems' behavior. Gaining a fundamental understanding of the nature and impact of such attacks will hopefully lead to more secure and robust social Web applications. We present the dimensions that characterize an attack and outline a framework to model the attacks based on various navigation channels and target elements. Using our framework we classify and identify different types of potential attack strategies against a social tagging system. We implement two of our attack models and evaluate their impact on retrieval algorithms commonly used by tagging systems.

Subjects: 12. Machine Learning and Discovery; 6.3 User Interfaces

Submitted: May 8, 2008

This page is copyrighted by AAAI. All rights reserved. Your use of this site constitutes acceptance of all of AAAI's terms and conditions and privacy policy.