Saved in:
Bibliographic Details
Main Authors: Martínez-Martínez, Josué, Brown, Olivia, Zeno, Giselle, Khorrami, Pooya, Caceres, Rajmonda
Format: Preprint
Published: 2026
Subjects:
Online Access:https://arxiv.org/abs/2601.09153
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866908763849490432
author Martínez-Martínez, Josué
Brown, Olivia
Zeno, Giselle
Khorrami, Pooya
Caceres, Rajmonda
author_facet Martínez-Martínez, Josué
Brown, Olivia
Zeno, Giselle
Khorrami, Pooya
Caceres, Rajmonda
contents Robustness to natural corruptions remains a critical challenge for reliable deep learning, particularly in safety-sensitive domains. We study a family of model-based training approaches that leverage a learned nuisance variation model to generate realistic corruptions, as well as new hybrid strategies that combine random coverage with adversarial refinement in nuisance space. Using the Challenging Unreal and Real Environments for Traffic Sign Recognition dataset (CURE-TSR), with Snow and Rain corruptions, we evaluate accuracy, calibration, and training complexity across corruption severities. Our results show that model-based methods consistently outperform baselines Vanilla, Adversarial Training, and AugMix baselines, with model-based adversarial training providing the strongest robustness under across all corruptions but at the expense of higher computation and model-based data augmentation achieving comparable robustness with $T$ less computational complexity without incurring a statistically significant drop in performance. These findings highlight the importance of learned nuisance models for capturing natural variability, and suggest a promising path toward more resilient and calibrated models under challenging conditions.
format Preprint
id arxiv_https___arxiv_org_abs_2601_09153
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle From Snow to Rain: Evaluating Robustness, Calibration, and Complexity of Model-Based Robust Training
Martínez-Martínez, Josué
Brown, Olivia
Zeno, Giselle
Khorrami, Pooya
Caceres, Rajmonda
Computer Vision and Pattern Recognition
Robustness to natural corruptions remains a critical challenge for reliable deep learning, particularly in safety-sensitive domains. We study a family of model-based training approaches that leverage a learned nuisance variation model to generate realistic corruptions, as well as new hybrid strategies that combine random coverage with adversarial refinement in nuisance space. Using the Challenging Unreal and Real Environments for Traffic Sign Recognition dataset (CURE-TSR), with Snow and Rain corruptions, we evaluate accuracy, calibration, and training complexity across corruption severities. Our results show that model-based methods consistently outperform baselines Vanilla, Adversarial Training, and AugMix baselines, with model-based adversarial training providing the strongest robustness under across all corruptions but at the expense of higher computation and model-based data augmentation achieving comparable robustness with $T$ less computational complexity without incurring a statistically significant drop in performance. These findings highlight the importance of learned nuisance models for capturing natural variability, and suggest a promising path toward more resilient and calibrated models under challenging conditions.
title From Snow to Rain: Evaluating Robustness, Calibration, and Complexity of Model-Based Robust Training
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2601.09153