Differentially private recommender systems

We consider the problem of producing recommendations from collective user behavior while simultaneously providing guarantees of privacy for these users. Specifically, we consider the Netflix Prize data set, and its leading algorithms, adapted to the framework of differential privacy.