Masks: Bringing Anonymity and Personalization Together


Prof. Virgilio Augusto F. Almeida

Universidade Federal de Minas Gerais


Dia 12/08/2003 das 10:30 às 11:30 – Sala H-301


Abstract - Most Web users realize that sites are collecting information about them, though few realize how much data is gathered or how that gathering occurs. Although some companies publish privacy policies to inform users about their practices, most policy statements are full of technical and legal jargon and are difficult to understand. Further, the Pew Internet & American Life Project's privacy survey revealed that, although US users are anxious about having their activities monitored, only 5 percent use tools to "anonymize" their requests, and only 10 percent reject cookies. Still other surveys (see, for example, doubleclicksummary.html) have found that users have a strong desire for personalization, which enhances services by customizing sites to users needs—but also requires sites to gather data to that end.

To bridge these conflicting needs, we developed Managing Anonymity while Sharing Knowledge to Servers, a Web-based framework that balances users' privacy concerns during Web browsing activities with their desire for personalized Web services. Masks uses a selective revelation scheme that erects an anonymity barrier between the user's private data and Web services, and controls the information that flows across that barrier to the service. This kind of filtering minimizes user information exposure while still permitting some form of service personalization. Also, because it addresses privacy at the data-collection level, Masks prevents third parties from building user profiles based on links to sites and information that might reveal personal information, such as religion, travel preferences, sexual orientation, and so on.



O Prof. Virgilio é professor e chefe do Departamento de Ciência da Computação da Universidade Federal de Minas Gerais e Ph.D. em Ciência da Computação pela Vanderbilt University. Virgílio ocupou os cargos de professor visitante na Boston University e pesquisador visitante na XEROX PARC e HP Research Labs. em Palo Alto, CA, EUA. Seus interesses em pesquisa são voltados para a área de modelagem e análise de desempenho em sistemas distribuídos em larga escala.