Collaborative Filtering Under a Sybil Attack: Similarity Metrics do Matter!

Abstract

Abstract-Recommendation systems help users identify interesting content, but they also open new privacy threats. In this paper, we deeply analyze the effect of a Sybil attack that tries to infer information on users from a user-based collaborativefiltering recommendation systems. We discuss the impact of different similarity metrics used to identity users with similar tastes in the trade-off between recommendation quality and privacy. Finally, we propose and evaluate a novel similarity metric that combines the best of both worlds: a high recommendation quality with a low prediction accuracy for the attacker. Our results, on a state-of-the-art recommendation framework and on real datasets show that existing similarity metrics exhibit a wide range of behaviors in the presence of Sybil attacks, while our new similarity metric consistently achieves the best trade-off while outperforming state-of-the-art solutions.

P3LS: Plausible Deniability for Practical Privacy-Preserving Live Streaming

“…Then, t-closeness [22] is a further extension of l-diversity. Instead of just guaranteeing a good representation of sensitive values, this approach enforces that the distribution of every sensitive attribute inside anonymity groups must be the same than the distribution of this attribute in the whole dataset, modulo a threshold t. b) Privacy in P2P: Several works have studied privacy in P2P systems, including privacy-preserving collaborative filtering [28], [29], accountability verifications [30], publishsubscribe systems [7], [31], or content-based routing with Intel SGX [32]. In the context of collaborative filtering, [17] shows how nodes can maintain an obfuscated profile of their interests, and exchange their profiles to route messages to the nodes that are willing to receive them.…”