Predictive Uncertainty Estimation via Prior Networks

TL;DR

This paper introduces Prior Networks, a novel framework for explicitly modeling predictive uncertainty by parameterizing a prior over predictive distributions, improving the detection of out-of-distribution samples and misclassifications.

Contribution

Prior Networks provide a new approach to explicitly model distributional uncertainty, outperforming previous methods in identifying OOD samples and misclassifications.

Findings

Prior Networks outperform previous methods in OOD detection.

They can distinguish between data and distributional uncertainty.

Effective on MNIST and CIFAR-10 datasets.

Abstract

Estimating how uncertain an AI system is in its predictions is important to improve the safety of such systems. Uncertainty in predictive can result from uncertainty in model parameters, irreducible data uncertainty and uncertainty due to distributional mismatch between the test and training data distributions. Different actions might be taken depending on the source of the uncertainty so it is important to be able to distinguish between them. Recently, baseline tasks and metrics have been defined and several practical methods to estimate uncertainty developed. These methods, however, attempt to model uncertainty due to distributional mismatch either implicitly through model uncertainty or as data uncertainty. This work proposes a new framework for modeling predictive uncertainty called Prior Networks (PNs) which explicitly models distributional uncertainty. PNs do this by parameterizing a prior distribution over predictive distributions. This work focuses on uncertainty for classification and evaluates PNs on the tasks of identifying out-of-distribution (OOD) samples and detecting misclassification on the MNIST dataset, where they are found to outperform previous methods. Experiments on synthetic and MNIST and CIFAR-10 data show that unlike previous non-Bayesian methods PNs are able to distinguish between data and distributional uncertainty.