On the Calibration of Image Semi-Supervised Learning Models
Proceedings of the AAAI Conference on Artificial Intelligence · 14.03.2026 · 40 (30) · 25073-25081
DOI: 10.1609/aaai.v40i30.39696 ↗Kurz gefasst
Fachartikel in der Zeitschrift Proceedings of the AAAI Conference on Artificial Intelligence, online erschienen am 14.03.2026. Verfasst von Mehrab Mustafy Rahman und 3 weiteren Autor:innen.
Abstract (Originalsprache)
Semi-supervised learning (SSL) has demonstrated high performance in image classification tasks by effectively utilizing both labeled and unlabeled data. However, existing SSL methods often suffer from poor calibration, with models yielding overconfident predictions that misrepresent actual prediction likelihoods. Recently, neural networks trained with mixup that linearly interpolates random examples from the training set have shown better calibration in supervised settings. However, calibration of neural models remains under-explored in SSL settings. Although effective in supervised model calibration, random mixup of pseudolabels in SSL presents challenges due to the overconfidence and unreliability of pseudolabels. In this work, we introduce CalibrateMix, a targeted mixup-based approach that aims to improve the calibration of SSL models while maintaining or even improving their classification accuracy. Our method leverages training dynamics of labeled and unlabeled samples to identify ''easy-to-learn'' and ''hard-to-learn'' samples, which in turn are utilized in a targeted mixup of easy and hard samples. Experimental results across several benchmark datasets show that our method achieves lower expected calibration error (ECE) and superior accuracy compared to existing SSL approaches.
Zitieren
APA
Rahman, M. M., Mohan, J., Sosea, T., & Caragea, C. (2026). On the Calibration of Image Semi-Supervised Learning Models. Proceedings of the AAAI Conference on Artificial Intelligence, 40(30), 25073-25081. https://doi.org/10.1609/aaai.v40i30.39696
Vancouver
Rahman MM, Mohan J, Sosea T, Caragea C. On the Calibration of Image Semi-Supervised Learning Models. Proceedings of the AAAI Conference on Artificial Intelligence. 2026;40(30):25073-25081. doi:10.1609/aaai.v40i30.39696
BibTeX
@article{Rahman2026,
author = {Rahman, Mehrab Mustafy and Mohan, Jayanth and Sosea, Tiberiu and Caragea, Cornelia},
title = {On the Calibration of Image Semi-Supervised Learning Models},
journal = {Proceedings of the AAAI Conference on Artificial Intelligence},
year = {2026},
volume = {40},
number = {30},
pages = {25073--25081},
doi = {10.1609/aaai.v40i30.39696}
}
EUVIMED