Privacy-preserving Hybrid Peer-to-Peer Recommendation System Architecture Locality-Sensitive Hashing in Structured Overlay Network

摘要

Recommendation systems are widely used to mitigate the information overflow peculiar to current life. Most of the modern recommendation system approaches are centralized. Although the centralized recommendations have some significant advantages they also bear two primary disadvantages: the necessity for users to share their preferences and a single point of failure. In this paper, an architecture of a collaborative peer-to-peer recommendation system with limited preferences' disclosure is proposed. Privacy in the proposed design is provided by the fact that exact user preferences are never shared together with the user identity. To achieve that, the proposed architecture employs a locality-sensitive hashing of user preferences and an anonymized distributed hash table approach to peer-to-peer design.