Too Sharp, Too Sure: When Calibration Follows Curvature

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Morosini, Alessandro, Gjika, Matea, Poggio, Tomaso, Beneventano, Pierfrancesco
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910157478297600
author Morosini, Alessandro
Gjika, Matea
Poggio, Tomaso
Beneventano, Pierfrancesco
author_facet Morosini, Alessandro
Gjika, Matea
Poggio, Tomaso
Beneventano, Pierfrancesco
contents Modern neural networks can achieve high accuracy while remaining poorly calibrated, producing confidence estimates that do not match empirical correctness. Yet calibration is often treated as a post-hoc attribute. We take a different perspective: we study calibration as a training-time phenomenon on small vision tasks, and ask whether calibrated solutions can be obtained reliably by intervening on the training procedure. We identify a tight coupling between calibration, curvature, and margins during training of deep networks under multiple gradient-based methods. Empirically, Expected Calibration Error (ECE) closely tracks curvature-based sharpness throughout optimization. Mathematically, we show that both ECE and Gauss--Newton curvature are controlled, up to problem-specific constants, by the same margin-dependent exponential tail functional along the trajectory. Guided by this mechanism, we introduce a margin-aware training objective that explicitly targets robust-margin tails and local smoothness, yielding improved out-of-sample calibration across optimizers without sacrificing accuracy.
format Preprint
id arxiv_https___arxiv_org_abs_2604_20614
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Too Sharp, Too Sure: When Calibration Follows Curvature
Morosini, Alessandro
Gjika, Matea
Poggio, Tomaso
Beneventano, Pierfrancesco
Machine Learning
Dynamical Systems
Optimization and Control
Modern neural networks can achieve high accuracy while remaining poorly calibrated, producing confidence estimates that do not match empirical correctness. Yet calibration is often treated as a post-hoc attribute. We take a different perspective: we study calibration as a training-time phenomenon on small vision tasks, and ask whether calibrated solutions can be obtained reliably by intervening on the training procedure. We identify a tight coupling between calibration, curvature, and margins during training of deep networks under multiple gradient-based methods. Empirically, Expected Calibration Error (ECE) closely tracks curvature-based sharpness throughout optimization. Mathematically, we show that both ECE and Gauss--Newton curvature are controlled, up to problem-specific constants, by the same margin-dependent exponential tail functional along the trajectory. Guided by this mechanism, we introduce a margin-aware training objective that explicitly targets robust-margin tails and local smoothness, yielding improved out-of-sample calibration across optimizers without sacrificing accuracy.
title Too Sharp, Too Sure: When Calibration Follows Curvature
topic Machine Learning
Dynamical Systems
Optimization and Control
url https://arxiv.org/abs/2604.20614