The Deep Weight Prior

Andrei Atanov, Arsenii Ashukha, Kirill Struminsky, Dmitriy Vetrov, Max Welling

Sep 27, 2018 ICLR 2019 Conference Blind Submission readers: everyone Show Bibtex
  • Abstract: Bayesian inference is known to provide a general framework for incorporating prior knowledge or specific properties into machine learning models via carefully choosing a prior distribution. In this work, we propose a new type of prior distributions for convolutional neural networks, deep weight prior, that in contrast to previously published techniques, favors empirically estimated structure of convolutional filters e.g., spatial correlations of weights. We define deep weight prior as an implicit distribution and propose a method for variational inference with such type of implicit priors. In experiments, we show that deep weight priors can improve the performance of Bayesian neural networks on several problems when training data is limited. Also, we found that initialization of weights of conventional networks with samples from deep weight prior leads to faster training.
  • Keywords: deep learning, variational inference, prior distributions
  • TL;DR: An empirical prior for convolutional layers in Bayesian neural networks that improves learning on small datasets.
0 Replies