Web-based Inference Detection
7Google Tech Talk
February 13, 2009
ABSTRACT
Presented by Jessica Staddon.
Text content can allow unintended inferences. Consider, for example, the numerous people who have published anonymous blogs for venting about their employer only to be identified through seemingly non-identifying posts. Similarly, the US government's "Operation Iraqi Freedom Portal" was assembled as evidence of nuclear weapons presence in Iraq, but removed because it could be used to infer much of the weapon making process. We propose a simple, semi-automated approach to detecting text-based inferences prior to the release of content. Our approach uses association rule mining of the Web to identify keywords that may allow a sensitive topic to be inferred. While the main application of this work is data leak prevention we will also discuss how it might be used to detect bias in product reviews. Finally, if time permits, we will discuss how inference detection can support topic-driven access control.
Most of this talk is joint work with Richard Chow and Philippe Golle.
Jessica is an area manager at PARC (aka Xerox PARC). She received her PhD in Math from U. C. Berkeley and has held research scientist positions at RSA Labs and Bell Labs. Jessica's background is in applied cryptography, specifically, cryptographic protocols for large, dynamic groups. Her current research interests include the use of data mining to support content privacy.
http://www.parc.com/jstaddon