ICML 2017poster593 citations
Multiplicative Normalizing Flows for Variational Bayesian Neural Networks
Abstract
We reinterpret multiplicative noise in neural networks as auxiliary random variables that augment the approximate posterior in a variational setting for Bayesian neural networks. We show that through this interpretation it is both efficient and straightforward to improve the approximation by employing normalizing flows while still allowing for local reparametrizations and a tractable lower bound. In experiments we show that with this new approximation we can significantly improve upon classical mean field for Bayesian neural networks on both predictive accuracy as well as predictive uncertainty.
BibTeX
@InProceedings{pmlr-v70-louizos17a,
title = {Multiplicative Normalizing Flows for Variational {B}ayesian Neural Networks},
author = {Christos Louizos and Max Welling},
booktitle = {Proceedings of the 34th International Conference on Machine Learning},
pages = {2218--2227},
year = {2017},
editor = {Precup, Doina and Teh, Yee Whye},
volume = {70},
series = {Proceedings of Machine Learning Research},
month = {06--11 Aug},
publisher = {PMLR},
pdf = {http://proceedings.mlr.press/v70/louizos17a/louizos17a.pdf},
url = {https://proceedings.mlr.press/v70/louizos17a.html},
abstract = {We reinterpret multiplicative noise in neural networks as auxiliary random variables that augment the approximate posterior in a variational setting for Bayesian neural networks. We show that through this interpretation it is both efficient and straightforward to improve the approximation by employing normalizing flows while still allowing for local reparametrizations and a tractable lower bound. In experiments we show that with this new approximation we can significantly improve upon classical mean field for Bayesian neural networks on both predictive accuracy as well as predictive uncertainty.}
}