Uncertainty Quantification and Deep Ensembles

TL;DR

This paper investigates the calibration issues of deep learning models in low-data scenarios, revealing that deep ensembles and regularization techniques can affect calibration and proposing simple strategies to improve calibration accuracy.

Contribution

It demonstrates that standard deep-ensembles may worsen calibration in low-data regimes and proposes a post-averaging temperature scaling method to significantly improve calibration.

Findings

Deep ensembles do not always improve calibration in low-data settings.

Temperature scaling after ensemble averaging reduces calibration error.

Combining data augmentation, ensembling, and calibration requires careful trade-offs.

Abstract

Deep Learning methods are known to suffer from calibration issues: they typically produce over-confident estimates. These problems are exacerbated in the low data regime. Although the calibration of probabilistic models is well studied, calibrating extremely over-parametrized models in the low-data regime presents unique challenges. We show that deep-ensembles do not necessarily lead to improved calibration properties. In fact, we show that standard ensembling methods, when used in conjunction with modern techniques such as mixup regularization, can lead to less calibrated models. This text examines the interplay between three of the most simple and commonly used approaches to leverage deep learning when data is scarce: data-augmentation, ensembling, and post-processing calibration methods. Although standard ensembling techniques certainly help boost accuracy, we demonstrate that the calibration of deep ensembles relies on subtle trade-offs. We also find that calibration methods such as temperature scaling need to be slightly tweaked when used with deep-ensembles and, crucially, need to be executed after the averaging process. Our simulations indicate that this simple strategy can halve the Expected Calibration Error (ECE) on a range of benchmark classification problems compared to standard deep-ensembles in the low data regime.