Can mutually mistrusting users collaborate in a distributed systems ? The answer is yes. See following papers.
2004 ACM Electronic Commerce: asynchronous trust system
This paper introduces the notion that Byzantine participants cannot bring down a recommendation system. It shows that honest peers can still successfully cooperate to retrieve the best value.
2005 IEEE ICDCS: concurrent trust system
2005 ACM-SIAM SODA: collaborative filtering
2005 ACM-SPAA Learning preferences by collaborating with strangers
This paper shows a randomized algorithm that successfully reconstructs all values even in presence of (majority) of byzantine peers.
2006 ACM-SPAA Collaborating with strangers in the presence of noise
2007 ACM-SPAA Minimizing online regret against dynamic optimum
Learning and adaptive systems
Can one learn from past mistakes and predict an unpredictable future ? The answer is yes. See following papers.
ACM STOC 1996: how to pick a winner
ACM PODC 2003: learning best network path in presence of oblivious failures
IEEE Infocom 2005:best network path in presence of adaptive failures,
ACM STOC 2004: learning best network path with round-trip feedback
Some other useful papers and links
Can one combine learning with collaborative filtering ? The answer is yes. See following papers.
2005 ACM Computational Learning Theory: competitive collaborative learning
This material is based upon work supported by the National Science Foundation under Grant No. 0617883, 0515080, 0240551, 0311795
Any opinions, findings and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation (NSF).