Carannante, G., Bouaynaya, N.C. and Mihaylova, L. orcid.org/0000-0001-5856-2223 (2021) An enhanced particle filter for uncertainty quantification in neural networks. In: de Villiers, P., de Waal, A. and Gustafsson, F., (eds.) 2021 IEEE 24th International Conference on Information Fusion (FUSION). 24th International Conference on Information Fusion (Fusion 2021), 01-04 Nov 2021, Sun City, South Africa. Institute of Electrical and Electronics Engineers ISBN 9781665414272
Abstract
The brittleness of deep learning models is ailing their deployment in real-world applications, such as transportation and airport security. Most work focuses on developing accurate models that only deliver point estimates without further information on model uncertainty or confidence. Ideally, a learning model should compute the posterior predictive distribution, which contains all information about the model output. We cast the problem of density tracking in neural networks using Particle Filtering, a powerful class of numerical methods for the solution of optimal estimation problems in non-linear, nonGaussian systems. Particle filters are a powerful alternative to Markov chain Monte Carlo algorithms and enjoy established convergence and performance guarantees. In this paper, we advance a particle filtering framework for neural networks, where the predictive output is a distribution. The mean of this distribution serves as the point estimate decision and its variance provides the model confidence in the decision. Our framework shows increased robustness under noisy conditions. Additionally, the predictive variance increases monotonically with decreasing signal-to-noise ratio (SNR); thus reflecting a lower confidence or higher uncertainty. This paper serves as a pioneering proof-of-concept framework that will allow the development of a theoretical understanding of robust neural networks.
Metadata
Item Type: | Proceedings Paper |
---|---|
Authors/Creators: |
|
Editors: |
|
Copyright, Publisher and Additional Information: | © 2021 ISIF. Personal use of this material is permitted. Permission from IEEE must be obtained for all other users, including reprinting/ republishing this material for advertising or promotional purposes, creating new collective works for resale or redistribution to servers or lists, or reuse of any copyrighted components of this work in other works. Reproduced in accordance with the publisher's self-archiving policy. |
Keywords: | Bayesian Learning; Particle Filtering; Neural Networks; Uncertainty Quantification |
Dates: |
|
Institution: | The University of Sheffield |
Academic Units: | The University of Sheffield > Faculty of Engineering (Sheffield) > Department of Automatic Control and Systems Engineering (Sheffield) |
Funding Information: | Funder Grant number Engineering and Physical Sciences Research Council EP/T013265/1 |
Depositing User: | Symplectic Sheffield |
Date Deposited: | 01 Sep 2021 10:14 |
Last Modified: | 02 Dec 2022 01:13 |
Published Version: | https://ieeexplore.ieee.org/stamp/stamp.jsp?arnumb... |
Status: | Published |
Publisher: | Institute of Electrical and Electronics Engineers |
Refereed: | Yes |
Related URLs: | |
Open Archives Initiative ID (OAI ID): | oai:eprints.whiterose.ac.uk:177516 |