EUVIMEDEUVIMED entdecken
Open Access · CC BY-NC 4.0

Reliable fairness auditing with semi-supervised inference

Jianhui Gao, Jessica Gronsbell

Biometrics · 2026

DOI: 10.1093/biomtc/ujag142 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Biometrics, 2026. Verfasst von Jianhui Gao und 1 weiteren Autor:innen. Beteiligte Einrichtung: Department of Statistical Science, University of Toronto, Toronto, ON M7A 2S4, Canada. Hauptthemen laut MEDLINE: Machine Learning, Models, Statistical. Die Arbeit nennt 33 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY-NC 4.0). Gefördert von: NSERC, University of Toronto.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Machine learning (ML) models often exhibit bias that can exacerbate inequities in biomedical applications. Fairness auditing, the process of evaluating a model’s performance across subpopulations, is critical for identifying and mitigating these biases. However, audits typically rely on large volumes of labeled data, which are costly and labor-intensive to obtain. To address this challenge, we introduce Infairness, a unified framework for auditing a wide range of fairness criteria using semi-supervised inference. Our approach combines a small labeled dataset with a large unlabeled dataset by imputing missing outcomes via regression with carefully selected nonlinear basis functions. Through extensive theoretical and empirical analyses, we show that our proposed estimator is (1) robust to specification of the ML or imputation model and (2) substantially more efficient than supervised estimation based solely on the labeled data. In two real-world fairness audits using electronic health record and medical imaging data, Infairness reduces variance by 40 - 60% compared to supervised estimation, underscoring its value for reliable fairness auditing with limited labeled data.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Machine LearningModels, StatisticalHumansElectronic Health RecordsBiasAlgorithmsData Interpretation, Statistical

Stichwörter der Autor:innen

group fairnessmachine learningmissing datasemi-supervised inference

Autor:innen und Institutionen

  • Department of Statistical Science, University of Toronto, Toronto, ON M7A 2S4, Canada.

Förderung

  • NSERC
  • University of Toronto

Zitieren

APA

Gao, J., & Gronsbell, J. (2026). Reliable fairness auditing with semi-supervised inference. Biometrics. https://doi.org/10.1093/biomtc/ujag142

Vancouver

Gao J, Gronsbell J. Reliable fairness auditing with semi-supervised inference. Biometrics. 2026. doi:10.1093/biomtc/ujag142

BibTeX
@article{Gao2026,
  author = {Gao, Jianhui and Gronsbell, Jessica},
  title = {Reliable fairness auditing with semi-supervised inference},
  journal = {Biometrics},
  year = {2026},
  doi = {10.1093/biomtc/ujag142}
}
Kostenlos starten →