Fast Max-Margin Matrix Factorization with Data Augmentation.

Minjie Xu,Jun Zhu,Bo Zhang
2013-01-01
Abstract:Existing max-margin matrix factorization (M 3 F) methods either are computationally inefficient or need a model selection procedure to determine the number of latent factors. In this paper we present a probabilistic M 3 F model that admits a highly efficient Gibbs sampling algorithm through data augmentation. We further extend our approach to incorporate Bayesian nonparametrics and build accordingly a truncation-free nonparametric M 3 F model where the number of latent factors is literally unbounded and inferred from data. Empirical studies on two large real-world data sets verify the efficacy of our proposed methods.
What problem does this paper attempt to address?