Stable Localized Conformal Prediction via Transduction

Yinjie Min, Liuhua Peng, Changliang Zou
Proceedings of the 43rd International Conference on Machine Learning, PMLR 306:88964-88989, 2026.

Abstract

Existing evaluations of conformal prediction, such as prediction efficiency and test-conditional coverage, are defined in expectation over the calibration data. In practice, when only one calibration set of limited size is available, prediction sets often exhibit high variability in size, especially for methods with localization. We formalize this concern as set stability, defined as the variance of the conditional expectation of the set size given the calibration data. To improve stability without requiring additional target-task labels, we propose Stable Conformal Prediction (StCP), a transfer learning approach that utilizes labeled source-task data and unlabeled target data. Theoretically, we characterize the marginal coverage and stability of StCP; empirically, it delivers more stable prediction sets than standard conformal prediction methods, especially for those with localization, when calibration data are limited.

Cite this Paper


BibTeX
@InProceedings{pmlr-v306-min26d, title = {Stable Localized Conformal Prediction via Transduction}, author = {Min, Yinjie and Peng, Liuhua and Zou, Changliang}, booktitle = {Proceedings of the 43rd International Conference on Machine Learning}, pages = {88964--88989}, year = {2026}, editor = {Zhang, Tong and Dudik, Miroslav and Jaggi, Martin and Agarwal, Alekh and Li, Sharon and Schuurmans, Dale and Zhu, Jerry and Berkenkamp, Felix and Dong, Hanze and Bietti, Alberto}, volume = {306}, series = {Proceedings of Machine Learning Research}, month = {06--11 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v306/main/assets/min26d/min26d.pdf}, url = {https://proceedings.mlr.press/v306/min26d.html}, abstract = {Existing evaluations of conformal prediction, such as prediction efficiency and test-conditional coverage, are defined in expectation over the calibration data. In practice, when only one calibration set of limited size is available, prediction sets often exhibit high variability in size, especially for methods with localization. We formalize this concern as set stability, defined as the variance of the conditional expectation of the set size given the calibration data. To improve stability without requiring additional target-task labels, we propose Stable Conformal Prediction (StCP), a transfer learning approach that utilizes labeled source-task data and unlabeled target data. Theoretically, we characterize the marginal coverage and stability of StCP; empirically, it delivers more stable prediction sets than standard conformal prediction methods, especially for those with localization, when calibration data are limited.} }
Endnote
%0 Conference Paper %T Stable Localized Conformal Prediction via Transduction %A Yinjie Min %A Liuhua Peng %A Changliang Zou %B Proceedings of the 43rd International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2026 %E Tong Zhang %E Miroslav Dudik %E Martin Jaggi %E Alekh Agarwal %E Sharon Li %E Dale Schuurmans %E Jerry Zhu %E Felix Berkenkamp %E Hanze Dong %E Alberto Bietti %F pmlr-v306-min26d %I PMLR %P 88964--88989 %U https://proceedings.mlr.press/v306/min26d.html %V 306 %X Existing evaluations of conformal prediction, such as prediction efficiency and test-conditional coverage, are defined in expectation over the calibration data. In practice, when only one calibration set of limited size is available, prediction sets often exhibit high variability in size, especially for methods with localization. We formalize this concern as set stability, defined as the variance of the conditional expectation of the set size given the calibration data. To improve stability without requiring additional target-task labels, we propose Stable Conformal Prediction (StCP), a transfer learning approach that utilizes labeled source-task data and unlabeled target data. Theoretically, we characterize the marginal coverage and stability of StCP; empirically, it delivers more stable prediction sets than standard conformal prediction methods, especially for those with localization, when calibration data are limited.
APA
Min, Y., Peng, L. & Zou, C.. (2026). Stable Localized Conformal Prediction via Transduction. Proceedings of the 43rd International Conference on Machine Learning, in Proceedings of Machine Learning Research 306:88964-88989 Available from https://proceedings.mlr.press/v306/min26d.html.

Related Material