This deliverable introduces privacy- and anti-discrimination by-design solutions for publishing and mining of targeted advertising datasets. The deliverable is divided into three building blocks: The first one presents the design of privacy analysis and data anonymization methods for preserving endusers’ individual privacy when publishing information-rich social networks. The second one presents the design of a holistic new framework for discovering and measuring different types of discrimination in a dataset of end-users’ historical decision records. The third one presents the design of privacy reserving recommended systems that are non-discriminatory by-design.
Please access the document in its entirety here.