Privacy preserving cooperative computation for personalized web search applications

Abstract

With the emergence of connected objects and the development of Artificial Intelligence (AI) mechanisms and algorithms, personalized applications are gaining an expanding interest, providing services tailored to each single user needs and expectations. They mainly rely on the massive collection of personal data generated by a large number of applications hosted from different connected devices. In this paper, we present CoWSA, a privacy preserving Cooperative computation framework for personalized Web Search peripheral Applications. The proposed framework is multi-fold. First, it provides the empowerment to end-users to control the disclosed personal data to third parties, while leveraging the trade-off between privacy and utility. Second, as a decentralized solution, CoWSA mitigates single points of failures, while ensuring the security of queries, the anonymity of submitting users, and the incentive of contributing nodes. Third, CoWSA is scalable as it provides acceptable computation and communication costs compared to most closely related schemes.