English

Calibration through the Lens of Interpretability

Machine Learning 2024-12-03 v1

Abstract

Calibration is a frequently invoked concept when useful label probability estimates are required on top of classification accuracy. A calibrated model is a function whose values correctly reflect underlying label probabilities. Calibration in itself however does not imply classification accuracy, nor human interpretable estimates, nor is it straightforward to verify calibration from finite data. There is a plethora of evaluation metrics (and loss functions) that each assess a specific aspect of a calibration model. In this work, we initiate an axiomatic study of the notion of calibration. We catalogue desirable properties of calibrated models as well as corresponding evaluation metrics and analyze their feasibility and correspondences. We complement this analysis with an empirical evaluation, comparing common calibration methods to employing a simple, interpretable decision tree.

Keywords

Cite

@article{arxiv.2412.00943,
  title  = {Calibration through the Lens of Interpretability},
  author = {Alireza Torabian and Ruth Urner},
  journal= {arXiv preprint arXiv:2412.00943},
  year   = {2024}
}

Comments

Published in XAI 2024

R2 v1 2026-06-28T20:18:48.599Z