Saved in:
Bibliographic Details
Main Authors: Wilson, Joseph, van der Heide, Chris, Hodgkinson, Liam, Roosta, Fred
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2502.02870
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866917069938753536
author Wilson, Joseph
van der Heide, Chris
Hodgkinson, Liam
Roosta, Fred
author_facet Wilson, Joseph
van der Heide, Chris
Hodgkinson, Liam
Roosta, Fred
contents While neural networks have demonstrated impressive performance across various tasks, accurately quantifying uncertainty in their predictions is essential to ensure their trustworthiness and enable widespread adoption in critical systems. Several Bayesian uncertainty quantification (UQ) methods exist that are either cheap or reliable, but not both. We propose a post-hoc, sampling-based UQ method for over-parameterized networks at the end of training. Our approach constructs efficient and meaningful deep ensembles by employing a (stochastic) gradient-descent sampling process on appropriately linearized networks. We demonstrate that our method effectively approximates the posterior of a Gaussian process using the empirical Neural Tangent Kernel. Through a series of numerical experiments, we show that our method not only outperforms competing approaches in computational efficiency-often reducing costs by multiple factors-but also maintains state-of-the-art performance across a variety of UQ metrics for both regression and classification tasks.
format Preprint
id arxiv_https___arxiv_org_abs_2502_02870
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Uncertainty Quantification with the Empirical Neural Tangent Kernel
Wilson, Joseph
van der Heide, Chris
Hodgkinson, Liam
Roosta, Fred
Machine Learning
While neural networks have demonstrated impressive performance across various tasks, accurately quantifying uncertainty in their predictions is essential to ensure their trustworthiness and enable widespread adoption in critical systems. Several Bayesian uncertainty quantification (UQ) methods exist that are either cheap or reliable, but not both. We propose a post-hoc, sampling-based UQ method for over-parameterized networks at the end of training. Our approach constructs efficient and meaningful deep ensembles by employing a (stochastic) gradient-descent sampling process on appropriately linearized networks. We demonstrate that our method effectively approximates the posterior of a Gaussian process using the empirical Neural Tangent Kernel. Through a series of numerical experiments, we show that our method not only outperforms competing approaches in computational efficiency-often reducing costs by multiple factors-but also maintains state-of-the-art performance across a variety of UQ metrics for both regression and classification tasks.
title Uncertainty Quantification with the Empirical Neural Tangent Kernel
topic Machine Learning
url https://arxiv.org/abs/2502.02870