Beyond One-Hot Labels: Semantic Mixing for Model Calibration

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Luo, Haoyang, Tao, Linwei, Dong, Minjing, Xu, Chang
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866916758594519040
author Luo, Haoyang
Tao, Linwei
Dong, Minjing
Xu, Chang
author_facet Luo, Haoyang
Tao, Linwei
Dong, Minjing
Xu, Chang
contents Model calibration seeks to ensure that models produce confidence scores that accurately reflect the true likelihood of their predictions being correct. However, existing calibration approaches are fundamentally tied to datasets of one-hot labels implicitly assuming full certainty in all the annotations. Such datasets are effective for classification but provides insufficient knowledge of uncertainty for model calibration, necessitating the curation of datasets with numerically rich ground-truth confidence values. However, due to the scarcity of uncertain visual examples, such samples are not easily available as real datasets. In this paper, we introduce calibration-aware data augmentation to create synthetic datasets of diverse samples and their ground-truth uncertainty. Specifically, we present \textbf{Calibration-aware Semantic Mixing (CSM)}, a novel framework that generates training samples with mixed class characteristics and annotates them with distinct confidence scores via diffusion models. Based on this framework, we propose calibrated reannotation to tackle the misalignment between the annotated confidence score and the mixing ratio during the diffusion reverse process. Besides, we explore the loss functions that better fit the new data representation paradigm. Experimental results demonstrate that CSM achieves superior calibration compared to the state-of-the-art calibration approaches. Our code is \href{https://github.com/E-Galois/CSM}{available here}.
format Preprint
id arxiv_https___arxiv_org_abs_2504_13548
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Beyond One-Hot Labels: Semantic Mixing for Model Calibration
Luo, Haoyang
Tao, Linwei
Dong, Minjing
Xu, Chang
Computer Vision and Pattern Recognition
Artificial Intelligence
Model calibration seeks to ensure that models produce confidence scores that accurately reflect the true likelihood of their predictions being correct. However, existing calibration approaches are fundamentally tied to datasets of one-hot labels implicitly assuming full certainty in all the annotations. Such datasets are effective for classification but provides insufficient knowledge of uncertainty for model calibration, necessitating the curation of datasets with numerically rich ground-truth confidence values. However, due to the scarcity of uncertain visual examples, such samples are not easily available as real datasets. In this paper, we introduce calibration-aware data augmentation to create synthetic datasets of diverse samples and their ground-truth uncertainty. Specifically, we present \textbf{Calibration-aware Semantic Mixing (CSM)}, a novel framework that generates training samples with mixed class characteristics and annotates them with distinct confidence scores via diffusion models. Based on this framework, we propose calibrated reannotation to tackle the misalignment between the annotated confidence score and the mixing ratio during the diffusion reverse process. Besides, we explore the loss functions that better fit the new data representation paradigm. Experimental results demonstrate that CSM achieves superior calibration compared to the state-of-the-art calibration approaches. Our code is \href{https://github.com/E-Galois/CSM}{available here}.
title Beyond One-Hot Labels: Semantic Mixing for Model Calibration
topic Computer Vision and Pattern Recognition
Artificial Intelligence
url https://arxiv.org/abs/2504.13548