Parallel and Distributed Processing Symposium, International (2013)
Cambridge, MA, USA USA
May 20, 2013 to May 24, 2013
DOI Bookmark: http://doi.ieeecomputersociety.org/10.1109/IPDPS.2013.47
We present WHATSUP, a collaborative filtering system for disseminating news items in a large-scale dynamic setting with no central authority. W HATS U P constructs an implicit social network based on user profiles that express the opinions of users about the news items they receive (like-dislike). Users with similar tastes are clustered using a similarity metric reflecting long-standing and emerging (dis)interests. News items are disseminated through a novel heterogeneous gossip protocol that (1) biases the orientation of its targets towards those with similar interests, and (2) amplifies dissemination based on the level of interest in every news item. We report on an extensive evaluation of W HATS U P through(a) simulations, (b) a Model Net emulation on a cluster, and(c) a Planet Lab deployment based on real datasets. We show that W HATS U P outperforms various alternatives in terms of accurate and complete delivery of relevant news items while preserving the fundamental advantages of standard gossip: namely, simplicity of deployment and robustness.
Protocols, Social network services, Measurement, Collaboration, Network topology, Communities, Accuracy, epidemic protocols, recommendation system, social networks
A. Boutet, D. Frey, R. Guerraoui, A. Jegou and A. Kermarrec, "WHATSUP: A Decentralized Instant News Recommender," 2013 IEEE 27th International Symposium on Parallel and Distributed Processing (IPDPS 2013)(IPDPS), Boston, MA, 2013, pp. 741-752.