Gaussian Mean Field Regularizes by Limiting Learned Information [article]

Julius Kunze, Louis Kirsch, Hippolyt Ritter, David Barber
2019 arXiv   pre-print
Variational inference with a factorized Gaussian posterior estimate is a widely used approach for learning parameters and hidden variables. Empirically, a regularizing effect can be observed that is poorly understood. In this work, we show how mean field inference improves generalization by limiting mutual information between learned parameters and the data through noise. We quantify a maximum capacity when the posterior variance is either fixed or learned and connect it to generalization
more » ... even when the KL-divergence in the objective is rescaled. Our experiments demonstrate that bounding information between parameters and data effectively regularizes neural networks on both supervised and unsupervised tasks.
arXiv:1902.04340v1 fatcat:nn3phevt25gktiognm46ckzkt4