Addressing Deep Learning Model Calibration Using Evidential Neural Networks and Uncertainty-Aware Training

TL;DR

This paper investigates how evidential neural networks and uncertainty-aware training can improve the calibration of deep learning models, especially in complex medical imaging tasks, thereby enhancing clinician trust.

Contribution

It systematically compares the effects of evidential neural networks and uncertainty-aware training, alone and combined, on model calibration in medical imaging applications.

Findings

Both methods improve calibration in complex tasks

Combining ENNs and uncertainty-aware training yields the best calibration results

Calibration issues worsen with higher model capacity in challenging tasks

Abstract

In terms of accuracy, deep learning (DL) models have had considerable success in classification problems for medical imaging applications. However, it is well-known that the outputs of such models, which typically utilise the SoftMax function in the final classification layer can be over-confident, i.e. they are poorly calibrated. Two competing solutions to this problem have been proposed: uncertainty-aware training and evidential neural networks (ENNs). In this paper, we perform an investigation into the improvements to model calibration that can be achieved by each of these approaches individually, and their combination. We perform experiments on two classification tasks: a simpler MNIST digit classification task and a more complex and realistic medical imaging artefact detection task using Phase Contrast Cardiac Magnetic Resonance images. The experimental results demonstrate that model calibration can suffer when the task becomes challenging enough to require a higher-capacity model. However, in our complex artefact detection task, we saw an improvement in calibration for both a low and higher-capacity model when implementing both the ENN and uncertainty-aware training together, indicating that this approach can offer a promising way to improve calibration in such settings. The findings highlight the potential use of these approaches to improve model calibration in a complex application, which would in turn improve clinician trust in DL models.