Prediction-Powered Risk Monitoring of Deployed Models for Detecting Harmful Distribution Shifts

Guangyi Zhang, Yunlong Cai, Guanding Yu, Osvaldo Simeone
Proceedings of the 43rd International Conference on Machine Learning, PMLR 306:154241-154262, 2026.

Abstract

We study the problem of monitoring model performance in dynamic environments where labeled data are limited. To this end, we propose prediction-powered risk monitoring (PPRM), a semi-supervised risk-monitoring approach based on prediction-powered inference (PPI). PPRM constructs anytime-valid lower bounds on the running risk by combining synthetic labels with a small set of true labels. Harmful shifts are detected via a threshold-based comparison with an upper bound on the nominal risk, satisfying assumption-free finite-sample guarantees on the type-I error. We demonstrate the effectiveness of PPRM through extensive experiments on image classification, large language model (LLM), and telecommunications monitoring tasks.

Cite this Paper

BibTeX
@InProceedings{pmlr-v306-zhang26g, title = {Prediction-Powered Risk Monitoring of Deployed Models for Detecting Harmful Distribution Shifts}, author = {Zhang, Guangyi and Cai, Yunlong and Yu, Guanding and Simeone, Osvaldo}, booktitle = {Proceedings of the 43rd International Conference on Machine Learning}, pages = {154241--154262}, 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/zhang26g/zhang26g.pdf}, url = {https://proceedings.mlr.press/v306/zhang26g.html}, abstract = {We study the problem of monitoring model performance in dynamic environments where labeled data are limited. To this end, we propose prediction-powered risk monitoring (PPRM), a semi-supervised risk-monitoring approach based on prediction-powered inference (PPI). PPRM constructs anytime-valid lower bounds on the running risk by combining synthetic labels with a small set of true labels. Harmful shifts are detected via a threshold-based comparison with an upper bound on the nominal risk, satisfying assumption-free finite-sample guarantees on the type-I error. We demonstrate the effectiveness of PPRM through extensive experiments on image classification, large language model (LLM), and telecommunications monitoring tasks.} }
Endnote
%0 Conference Paper %T Prediction-Powered Risk Monitoring of Deployed Models for Detecting Harmful Distribution Shifts %A Guangyi Zhang %A Yunlong Cai %A Guanding Yu %A Osvaldo Simeone %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-zhang26g %I PMLR %P 154241--154262 %U https://proceedings.mlr.press/v306/zhang26g.html %V 306 %X We study the problem of monitoring model performance in dynamic environments where labeled data are limited. To this end, we propose prediction-powered risk monitoring (PPRM), a semi-supervised risk-monitoring approach based on prediction-powered inference (PPI). PPRM constructs anytime-valid lower bounds on the running risk by combining synthetic labels with a small set of true labels. Harmful shifts are detected via a threshold-based comparison with an upper bound on the nominal risk, satisfying assumption-free finite-sample guarantees on the type-I error. We demonstrate the effectiveness of PPRM through extensive experiments on image classification, large language model (LLM), and telecommunications monitoring tasks.
APA
Zhang, G., Cai, Y., Yu, G. & Simeone, O.. (2026). Prediction-Powered Risk Monitoring of Deployed Models for Detecting Harmful Distribution Shifts. Proceedings of the 43rd International Conference on Machine Learning, in Proceedings of Machine Learning Research 306:154241-154262 Available from https://proceedings.mlr.press/v306/zhang26g.html.

Related Material