Quantum recommendation systems

Qu&Co comments on this publication:

A recommendation engine uses the past purchases or ratings to construct a (partial) preference matrix, which is used to provide personalized recommendations to individual users. In this paper, Kerenidis et al. present a quantum recommendation system which updates the partial preference matrix each time data comes in and can, based on such matrix, provides a recommendation in time polylog to the matrix-dimension, exponentially faster than classical recommendation systems.