Asynchronous recommendation systems

Baruch Awerbuch*, Aviv Nisgav, Boaz Patt-Shamir

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

Motivation. Recommendation systems are an important ingredient of modern life, where people must make decisions with partial information [6]. Everyday examples include buying books, going to a movie, choosing an on-line store etc. Computer-related examples include, among others, choosing peers in a potentially hostile peer-to-peer environment, or choosing a route in an unreliable network. The basic idea underlying such systems is that users can use the experience reported by others so as to improve their prediction of their own opinions. However, users may differ in their opinions either because they have different tastes, or because their objectives may be different (e.g., in a peer-to-peer network some users may wish to destroy the system). Obviously, only users who share preferences (said to belong to the same type) are can enjoy the advantages of recommendation systems.

Original languageEnglish
Title of host publicationPODC'07
Subtitle of host publicationProceedings of the 26th Annual ACM Symposium on Principles of Distributed Computing
Pages366-367
Number of pages2
DOIs
StatePublished - 2007
EventPODC'07: 26th Annual ACM Symposium on Principles of Distributed Computing - Portland, OR, United States
Duration: 12 Aug 200715 Aug 2007

Publication series

NameProceedings of the Annual ACM Symposium on Principles of Distributed Computing

Conference

ConferencePODC'07: 26th Annual ACM Symposium on Principles of Distributed Computing
Country/TerritoryUnited States
CityPortland, OR
Period12/08/0715/08/07

Keywords

  • Asynchronous distributed algorithms
  • Collaborative filtering
  • Recommender systems

Fingerprint

Dive into the research topics of 'Asynchronous recommendation systems'. Together they form a unique fingerprint.

Cite this