An EP algorithm for learning highly interpretable classifiers | Publicación