Toggle Main Menu Toggle Search

Open Access padlockePrints

Deep Bayesian Gaussian processes for uncertainty estimation in electronic health records

Lookup NU author(s): Dr Ali HassaineORCiD, Dr Dexter CanoyORCiD



This work is licensed under a Creative Commons Attribution 4.0 International License (CC BY 4.0).


© 2021, The Author(s).One major impediment to the wider use of deep learning for clinical decision making is the difficulty of assigning a level of confidence to model predictions. Currently, deep Bayesian neural networks and sparse Gaussian processes are the main two scalable uncertainty estimation methods. However, deep Bayesian neural networks suffer from lack of expressiveness, and more expressive models such as deep kernel learning, which is an extension of sparse Gaussian process, captures only the uncertainty from the higher-level latent space. Therefore, the deep learning model under it lacks interpretability and ignores uncertainty from the raw data. In this paper, we merge features of the deep Bayesian learning framework with deep kernel learning to leverage the strengths of both methods for a more comprehensive uncertainty estimation. Through a series of experiments on predicting the first incidence of heart failure, diabetes and depression applied to large-scale electronic medical records, we demonstrate that our method is better at capturing uncertainty than both Gaussian processes and deep Bayesian neural networks in terms of indicating data insufficiency and identifying misclassifications, with a comparable generalization performance. Furthermore, by assessing the accuracy and area under the receiver operating characteristic curve over the predictive probability, we show that our method is less susceptible to making overconfident predictions, especially for the minority class in imbalanced datasets. Finally, we demonstrate how uncertainty information derived by the model can inform risk factor analysis towards model interpretability.

Publication metadata

Author(s): Li Y, Rao S, Hassaine A, Ramakrishnan R, Canoy D, Salimi-Khorshidi G, Mamouei M, Lukasiewicz T, Rahimi K

Publication type: Article

Publication status: Published

Journal: Scientific Reports

Year: 2021

Volume: 11

Issue: 1

Online publication date: 19/10/2021

Acceptance date: 05/10/2021

Date deposited: 25/11/2022

ISSN (electronic): 2045-2322

Publisher: Nature Research


DOI: 10.1038/s41598-021-00144-6

PubMed id: 34667200


Altmetrics provided by Altmetric


Funder referenceFunder name
BHF KR & DC (grant number: PG/18/65/33872)
British Heart Foundation (BHF): YL & KR (grant number: FS/PhD/21/29110)
KR is also in receipt of funding from the UKRI’s Global Challenges Research Fund (GCRF), Grant Ref ES/P0110551/1,