A Utility-Theoretic Approach to Privacy and Personalization

Andreas Krause and Eric Horvitz

Access pdf.


Online services such as web search, news portals, and e-commerce applications face the challenge of providing high quality experiences to a large, heterogeneous user base. Recent efforts have highlighted the potential to improve performance by personalizing services based on special knowledge about users. For example, a userís location, demographics, and search and browsing history may be useful in enhancing the results offered in response to web search queries. However, reasonable concerns about privacy by users, providers, and government agencies acting on behalf of citizens, may limit access to such information. We introduce and explore an economics of privacy in personalization, where people can opt to share personal information in return for enhancements in the quality of an online service. We focus on the example of web search and formulate realistic objective functions for search efficacy and privacy. We demonstrate how we can identify a near-optimal solution to the utility privacy tradeoff. We evaluate the methodology on data drawn from a log of the search activity of volunteer participants. We separately assess usersí preferences about privacy and utility via a large-scale survey, aimed at eliciting preferences about peoplesí willingness to trade the sharing of personal data in returns for gains in search efficiency. We show that a significant level of personalization can be achieved using only a small amount of information about users.

Keywords: Utility theory, privacy, web search, personalization

In: Proceedings of AAAI-08: Twenty-Third Conference on Artificial Intelligence, Chicago, Illinois, July 2008.

Back to Eric Horvitz's home page.