RECOMMENDER SYSTEM AND MEDIA RETRIEVAL SYSTEM FOR PROVIDING RECOMMENDATIONS TO GROUPS OF USERS

摘要

The recommender system uses a processing system configured to determine recommendation scores using a collaborative rating process. Collaborative rating process involves determining correlations between ratings for media items provided by a current user for which a recommendation score has to be computed and by further users, and computing recommendation scores from ratings for other media items from ratings from the further users, dependent on the correlation. To handle groups, information is determined by identifying a plurality of users is determined that concurrently use a user terminal, and group correlation values for use in the collaborative rating process are computed from rating values from individual ones of the users in the group. Synthetic rating values for the group may be computed from a sum of rating values of the members and then used to compute the correlation. The computation of the group correlation is implemented using a user interface system that encrypts information the identity of the users and transmits that information to a processing system with a server and a privacy service provider system. The server and a privacy service provider system use a secure multi-party computation protocol to perform a distributed computation of the group correlation.