Saved in:
Bibliographic Details
Main Authors: Ciosek, Kamil, Felicioni, Nicolò, Ghiassian, Sina, Litwin, Juan Elenter, Tonolini, Francesco, Gustafsson, David, Garcia-Martin, Eva, Gonzalez, Carmen Barcena, Bertrand-Lalo, Raphaëlle
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2512.13872
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866908867729817600
author Ciosek, Kamil
Felicioni, Nicolò
Ghiassian, Sina
Litwin, Juan Elenter
Tonolini, Francesco
Gustafsson, David
Garcia-Martin, Eva
Gonzalez, Carmen Barcena
Bertrand-Lalo, Raphaëlle
author_facet Ciosek, Kamil
Felicioni, Nicolò
Ghiassian, Sina
Litwin, Juan Elenter
Tonolini, Francesco
Gustafsson, David
Garcia-Martin, Eva
Gonzalez, Carmen Barcena
Bertrand-Lalo, Raphaëlle
contents We make two contributions to the problem of estimating the $L_1$ calibration error of a binary classifier from a finite dataset. First, we provide an upper bound for any classifier where the calibration function has bounded variation. Second, we provide a method of modifying any classifier so that its calibration error can be upper bounded efficiently without significantly impacting classifier performance and without any restrictive assumptions. All our results are non-asymptotic and distribution-free. We conclude by providing advice on how to measure calibration error in practice. Our methods yield practical procedures that can be run on real-world datasets with modest overhead.
format Preprint
id arxiv_https___arxiv_org_abs_2512_13872
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Measuring Uncertainty Calibration
Ciosek, Kamil
Felicioni, Nicolò
Ghiassian, Sina
Litwin, Juan Elenter
Tonolini, Francesco
Gustafsson, David
Garcia-Martin, Eva
Gonzalez, Carmen Barcena
Bertrand-Lalo, Raphaëlle
Machine Learning
We make two contributions to the problem of estimating the $L_1$ calibration error of a binary classifier from a finite dataset. First, we provide an upper bound for any classifier where the calibration function has bounded variation. Second, we provide a method of modifying any classifier so that its calibration error can be upper bounded efficiently without significantly impacting classifier performance and without any restrictive assumptions. All our results are non-asymptotic and distribution-free. We conclude by providing advice on how to measure calibration error in practice. Our methods yield practical procedures that can be run on real-world datasets with modest overhead.
title Measuring Uncertainty Calibration
topic Machine Learning
url https://arxiv.org/abs/2512.13872