Enabling Uncertainty Estimation in Iterative Neural Networks

Nikita Durasov, Doruk Oner, Jonathan Donier, Hieu Le, Pascal Fua
Proceedings of the 41st International Conference on Machine Learning, PMLR 235:12172-12189, 2024.

Abstract

Turning pass-through network architectures into iterative ones, which use their own output as input, is a well-known approach for boosting performance. In this paper, we argue that such architectures offer an additional benefit: The convergence rate of their successive outputs is highly correlated with the accuracy of the value to which they converge. Thus, we can use the convergence rate as a useful proxy for uncertainty. This results in an approach to uncertainty estimation that provides state-of-the-art estimates at a much lower computational cost than techniques like Ensembles, and without requiring any modifications to the original iterative model. We demonstrate its practical value by embedding it in two application domains: road detection in aerial images and the estimation of aerodynamic properties of 2D and 3D shapes.

Cite this Paper

BibTeX
@InProceedings{pmlr-v235-durasov24a, title = {Enabling Uncertainty Estimation in Iterative Neural Networks}, author = {Durasov, Nikita and Oner, Doruk and Donier, Jonathan and Le, Hieu and Fua, Pascal}, booktitle = {Proceedings of the 41st International Conference on Machine Learning}, pages = {12172--12189}, year = {2024}, editor = {Salakhutdinov, Ruslan and Kolter, Zico and Heller, Katherine and Weller, Adrian and Oliver, Nuria and Scarlett, Jonathan and Berkenkamp, Felix}, volume = {235}, series = {Proceedings of Machine Learning Research}, month = {21--27 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v235/main/assets/durasov24a/durasov24a.pdf}, url = {https://proceedings.mlr.press/v235/durasov24a.html}, abstract = {Turning pass-through network architectures into iterative ones, which use their own output as input, is a well-known approach for boosting performance. In this paper, we argue that such architectures offer an additional benefit: The convergence rate of their successive outputs is highly correlated with the accuracy of the value to which they converge. Thus, we can use the convergence rate as a useful proxy for uncertainty. This results in an approach to uncertainty estimation that provides state-of-the-art estimates at a much lower computational cost than techniques like Ensembles, and without requiring any modifications to the original iterative model. We demonstrate its practical value by embedding it in two application domains: road detection in aerial images and the estimation of aerodynamic properties of 2D and 3D shapes.} }
Endnote
%0 Conference Paper %T Enabling Uncertainty Estimation in Iterative Neural Networks %A Nikita Durasov %A Doruk Oner %A Jonathan Donier %A Hieu Le %A Pascal Fua %B Proceedings of the 41st International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2024 %E Ruslan Salakhutdinov %E Zico Kolter %E Katherine Heller %E Adrian Weller %E Nuria Oliver %E Jonathan Scarlett %E Felix Berkenkamp %F pmlr-v235-durasov24a %I PMLR %P 12172--12189 %U https://proceedings.mlr.press/v235/durasov24a.html %V 235 %X Turning pass-through network architectures into iterative ones, which use their own output as input, is a well-known approach for boosting performance. In this paper, we argue that such architectures offer an additional benefit: The convergence rate of their successive outputs is highly correlated with the accuracy of the value to which they converge. Thus, we can use the convergence rate as a useful proxy for uncertainty. This results in an approach to uncertainty estimation that provides state-of-the-art estimates at a much lower computational cost than techniques like Ensembles, and without requiring any modifications to the original iterative model. We demonstrate its practical value by embedding it in two application domains: road detection in aerial images and the estimation of aerodynamic properties of 2D and 3D shapes.
APA
Durasov, N., Oner, D., Donier, J., Le, H. & Fua, P.. (2024). Enabling Uncertainty Estimation in Iterative Neural Networks. Proceedings of the 41st International Conference on Machine Learning, in Proceedings of Machine Learning Research 235:12172-12189 Available from https://proceedings.mlr.press/v235/durasov24a.html.

Related Material