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

Tareen Dawood*, Emily Chan, Reza Razavi, Andrew P. King, Esther Puyol-Anton

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingOther chapter contributionpeer-review

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.

Original languageEnglish
Title of host publication2023 IEEE International Symposium on Biomedical Imaging, ISBI 2023
PublisherIEEE Computer Society
ISBN (Electronic)9781665473583
DOIs
Publication statusPublished - 2023
Event20th IEEE International Symposium on Biomedical Imaging, ISBI 2023 - Cartagena, Colombia
Duration: 18 Apr 202321 Apr 2023

Publication series

NameProceedings - International Symposium on Biomedical Imaging
Volume2023-April
ISSN (Print)1945-7928
ISSN (Electronic)1945-8452

Conference

Conference20th IEEE International Symposium on Biomedical Imaging, ISBI 2023
Country/TerritoryColombia
CityCartagena
Period18/04/202321/04/2023

Keywords

  • Calibration
  • Evidential Neural Networks
  • Medical Imaging

Fingerprint

Dive into the research topics of 'Addressing Deep Learning Model Calibration Using Evidential Neural Networks and Uncertainty-Aware Training'. Together they form a unique fingerprint.

Cite this