Affiner les résultats

Gaussian mean field regularizes by limiting learned information

Kunze, Julius ; Kirsch, Louis ; Ritter, Hippolyt ; Barber, David

In: Entropy, 2019, vol. 21, no. 8, p. 758

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...