Noise Contrastive Priors for Functional Uncertainty

TL;DR

This paper introduces Noise Contrastive Priors (NCPs), a novel method for neural networks to produce reliable uncertainty estimates, especially outside the training distribution, enhancing active learning and scalability.

Contribution

The paper proposes NCPs, a new approach that trains models to output high uncertainty on out-of-distribution data, improving uncertainty estimation and scalability.

Findings

NCPs prevent overfitting outside training data

NCPs improve active learning performance

NCPs scale effectively to large datasets

Abstract

Obtaining reliable uncertainty estimates of neural network predictions is a long standing challenge. Bayesian neural networks have been proposed as a solution, but it remains open how to specify their prior. In particular, the common practice of an independent normal prior in weight space imposes relatively weak constraints on the function posterior, allowing it to generalize in unforeseen ways on inputs outside of the training distribution. We propose noise contrastive priors (NCPs) to obtain reliable uncertainty estimates. The key idea is to train the model to output high uncertainty for data points outside of the training distribution. NCPs do so using an input prior, which adds noise to the inputs of the current mini batch, and an output prior, which is a wide distribution given these inputs. NCPs are compatible with any model that can output uncertainty estimates, are easy to scale, and yield reliable uncertainty estimates throughout training. Empirically, we show that NCPs prevent overfitting outside of the training distribution and result in uncertainty estimates that are useful for active learning. We demonstrate the scalability of our method on the flight delays data set, where we significantly improve upon previously published results.