- PRIVACY-PRESERVING ADVERTISEMENT TARGETING USING RANDOMIZED PROFILE PERTURBATION

摘要

An advertisement distribution and scheduling system is provided that targets users to ads and maximizes service-provider revenue without a complete understanding of user-profile information. Each user device stores a user profile and is pre-loaded with a set of ads to be viewed during a timeslot. Each user device selects and displays an advertisement based on the user profile but does not verify the selected advertisement to the service provider. Instead, the user devices provide disturbed user-profile information in the form of Boolean vectors for use with guaranteed-approximation on-line algorithms to estimate the number of users who have viewed the particular advertisement by the service provider. Thus, the service provider can bill advertisers for the number of times their ads are watched without having to know about the profiles of the users or the advertisements viewed by the individual users, The targeted advertisements can be watched while keeping the privacy.