Bayesian Recurrent Neural Networks

TL;DR

This paper presents a simple variational Bayes approach for RNNs that improves uncertainty estimation, regularization, and performance, with novel posterior approximation techniques and broad applicability.

Contribution

It introduces a straightforward adaptation of truncated backpropagation through time for Bayesian RNNs and a new posterior approximation method that enhances model performance.

Findings

Bayesian RNNs outperform traditional RNNs on language modeling and image captioning.

The proposed methods reduce parameters by 80% while maintaining accuracy.

A new benchmark for uncertainty in language models is introduced.

Abstract

In this work we explore a straightforward variational Bayes scheme for Recurrent Neural Networks. Firstly, we show that a simple adaptation of truncated backpropagation through time can yield good quality uncertainty estimates and superior regularisation at only a small extra computational cost during training, also reducing the amount of parameters by 80\%. Secondly, we demonstrate how a novel kind of posterior approximation yields further improvements to the performance of Bayesian RNNs. We incorporate local gradient information into the approximate posterior to sharpen it around the current batch statistics. We show how this technique is not exclusive to recurrent neural networks and can be applied more widely to train Bayesian neural networks. We also empirically demonstrate how Bayesian RNNs are superior to traditional RNNs on a language modelling benchmark and an image captioning task, as well as showing how each of these methods improve our model over a variety of other schemes for training them. We also introduce a new benchmark for studying uncertainty for language models so future methods can be easily compared.