Recommending with an Agenda: Active Learning of Private Attributes using Matrix Factorization

Recommender systems leverage user social and demographic information, e.g., age, gender and political affiliation, to personalize content and monetize on their users. Oftentimes, users do not volunteer this information due to privacy concerns or to the lack of initiative in filling out their profile information. In this work, we illustrate a new threat in which the system can nevertheless learn the private attribute for those users who do not voluntarily disclose them. We design an active attack that solicit ratings for strategically selected items, and could thus be used by a recommender system to pursue its hidden agenda. Our method is based on a novel usage of Bayesian matrix factorization in an active learning setting. Evaluations, on multiple datasets, illustrate that such an attack is indeed feasible and can be carried out using significantly fewer rated items than the previously proposed static inference methods. Importantly, this threat can succeed without sacrificing the quality of the regular recommendations made to the user.