Section: Application Domains

News distribution and collaborative editing

We want to create a new application that allows users to distribute news using collaborative filtering. News would be gossipped among peers based on the rating assigned to news items by the various users. Furthermore, ratings would influence the timeline of news items displayed for each user, reflecting the user's preferences. A reputation system would enable established contributors to have their articles start with a higher a-priori ranking, allowing them to instantly rise above the noise generated by advertising. New contributors can use a proof-of-work calculation to increase the visibility of their work. The payment system can be used to reward contributors.

When peers compare scores, preserving the privacy of the individual rankings is important as users might not want to expose their political views, and as malicious participants might be able to game the process if they are able to determine the ranking of another peer. We thus propose to use the SMC scalar product (together with an efficient set intersection mechanism to deal with sparcity) for these joint computations.