Learn then test: Calibrating predictive algorithms to achieve risk control
The Annals of Applied Statistics · 01.06.2025 · 19 (2) · 1641-1662
DOI: 10.1214/24-aoas1998 ↗Kurz gefasst
Fachartikel in der Zeitschrift The Annals of Applied Statistics, erschienen am 01.06.2025. Verfasst von Anastasios N. Angelopoulos und 4 weiteren Autor:innen. Die Arbeit wurde bisher 54-mal zitiert (Crossref) und nennt 66 Literaturangaben.
Abstract (Originalsprache)
We introduce a framework for calibrating machine learning models to satisfy finite-sample statistical guarantees. Our calibration algorithms work with any model and (unknown) data-generating distribution and do not require retraining. The algorithms address, among other examples, false discovery rate control in multilabel classification, intersection-over-union control in instance segmentation, and simultaneous control of the type-1 outlier error and confidence set coverage in classification or regression. Our main insight is to reframe risk control as multiple hypothesis testing, enabling different mathematical arguments. We demonstrate our algorithms with detailed worked examples in computer vision and tabular medical data. The computer vision experiments demonstrate the utility of our approach in calibrating state-of-the-art predictive architectures that have been deployed widely, such as the detectron2 object detection system.
Zitieren
APA
Angelopoulos, A. N., Bates, S., Candès, E. J., Jordan, M. I., & Lei, L. (2025). Learn then test: Calibrating predictive algorithms to achieve risk control. The Annals of Applied Statistics, 19(2), 1641-1662. https://doi.org/10.1214/24-aoas1998
Vancouver
Angelopoulos AN, Bates S, Candès EJ, Jordan MI, Lei L. Learn then test: Calibrating predictive algorithms to achieve risk control. The Annals of Applied Statistics. 2025;19(2):1641-1662. doi:10.1214/24-aoas1998
BibTeX
@article{Angelopoulos2025,
author = {Angelopoulos, Anastasios N. and Bates, Stephen and Candès, Emmanuel J. and Jordan, Michael I. and Lei, Lihua},
title = {Learn then test: Calibrating predictive algorithms to achieve risk control},
journal = {The Annals of Applied Statistics},
year = {2025},
volume = {19},
number = {2},
pages = {1641--1662},
doi = {10.1214/24-aoas1998}
}
EUVIMED