Saved in:
Bibliographic Details
Main Authors: Kirchhof, Michael, Collier, Mark, Oh, Seong Joon, Kasneci, Enkelejda
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2402.16569
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866929257698033664
author Kirchhof, Michael
Collier, Mark
Oh, Seong Joon
Kasneci, Enkelejda
author_facet Kirchhof, Michael
Collier, Mark
Oh, Seong Joon
Kasneci, Enkelejda
contents Accurate uncertainty estimation is vital to trustworthy machine learning, yet uncertainties typically have to be learned for each task anew. This work introduces the first pretrained uncertainty modules for vision models. Similar to standard pretraining this enables the zero-shot transfer of uncertainties learned on a large pretraining dataset to specialized downstream datasets. We enable our large-scale pretraining on ImageNet-21k by solving a gradient conflict in previous uncertainty modules and accelerating the training by up to 180x. We find that the pretrained uncertainties generalize to unseen datasets. In scrutinizing the learned uncertainties, we find that they capture aleatoric uncertainty, disentangled from epistemic components. We demonstrate that this enables safe retrieval and uncertainty-aware dataset visualization. To encourage applications to further problems and domains, we release all pretrained checkpoints and code under https://github.com/mkirchhof/url .
format Preprint
id arxiv_https___arxiv_org_abs_2402_16569
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Pretrained Visual Uncertainties
Kirchhof, Michael
Collier, Mark
Oh, Seong Joon
Kasneci, Enkelejda
Computer Vision and Pattern Recognition
Machine Learning
Accurate uncertainty estimation is vital to trustworthy machine learning, yet uncertainties typically have to be learned for each task anew. This work introduces the first pretrained uncertainty modules for vision models. Similar to standard pretraining this enables the zero-shot transfer of uncertainties learned on a large pretraining dataset to specialized downstream datasets. We enable our large-scale pretraining on ImageNet-21k by solving a gradient conflict in previous uncertainty modules and accelerating the training by up to 180x. We find that the pretrained uncertainties generalize to unseen datasets. In scrutinizing the learned uncertainties, we find that they capture aleatoric uncertainty, disentangled from epistemic components. We demonstrate that this enables safe retrieval and uncertainty-aware dataset visualization. To encourage applications to further problems and domains, we release all pretrained checkpoints and code under https://github.com/mkirchhof/url .
title Pretrained Visual Uncertainties
topic Computer Vision and Pattern Recognition
Machine Learning
url https://arxiv.org/abs/2402.16569