CARE: Adaptive Calibration for Reliable Recommendations

Nitin Bisht, Huan Huo, Xiuwen Gong, Guandong Xu
Proceedings of the 43rd International Conference on Machine Learning, PMLR 306:8340-8372, 2026.

Abstract

Modern recommender systems are typically trained offline and deployed with parameters held fixed between periodic refreshes, yet user behavior can evolve substantially during deployment. This can cause ranking utility to degrade over time and makes it difficult to provide formal guarantees about recommendation quality. We propose CARE, an adaptive calibration framework that wraps an arbitrary backbone recommender and outputs variable-size recommendation sets with finite-sample performance guarantees over interaction streams. CARE combines (i) a loss-based monitoring module that localizes behavioral changes and triggers threshold recalibration, and (ii) an online aggregation rule that promotes compact recommendation sets by dynamically reweighting candidate set predictors. We provide theoretical results establishing finite-sample guarantees for utility-based risk control and bounds on the expected set size relative to the best constituent predictor. Experiments across multiple datasets and backbone models demonstrate that CARE improves robustness and maintains compact recommendation sets while preserving the desired statistical guarantees. The code and implementation are available in https://github.com/kalpiree/CARE.

Cite this Paper

BibTeX
@InProceedings{pmlr-v306-bisht26b, title = {{CARE}: Adaptive Calibration for Reliable Recommendations}, author = {Bisht, Nitin and Huo, Huan and Gong, Xiuwen and Xu, Guandong}, booktitle = {Proceedings of the 43rd International Conference on Machine Learning}, pages = {8340--8372}, 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/bisht26b/bisht26b.pdf}, url = {https://proceedings.mlr.press/v306/bisht26b.html}, abstract = {Modern recommender systems are typically trained offline and deployed with parameters held fixed between periodic refreshes, yet user behavior can evolve substantially during deployment. This can cause ranking utility to degrade over time and makes it difficult to provide formal guarantees about recommendation quality. We propose CARE, an adaptive calibration framework that wraps an arbitrary backbone recommender and outputs variable-size recommendation sets with finite-sample performance guarantees over interaction streams. CARE combines (i) a loss-based monitoring module that localizes behavioral changes and triggers threshold recalibration, and (ii) an online aggregation rule that promotes compact recommendation sets by dynamically reweighting candidate set predictors. We provide theoretical results establishing finite-sample guarantees for utility-based risk control and bounds on the expected set size relative to the best constituent predictor. Experiments across multiple datasets and backbone models demonstrate that CARE improves robustness and maintains compact recommendation sets while preserving the desired statistical guarantees. The code and implementation are available in https://github.com/kalpiree/CARE.} }
Endnote
%0 Conference Paper %T CARE: Adaptive Calibration for Reliable Recommendations %A Nitin Bisht %A Huan Huo %A Xiuwen Gong %A Guandong Xu %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-bisht26b %I PMLR %P 8340--8372 %U https://proceedings.mlr.press/v306/bisht26b.html %V 306 %X Modern recommender systems are typically trained offline and deployed with parameters held fixed between periodic refreshes, yet user behavior can evolve substantially during deployment. This can cause ranking utility to degrade over time and makes it difficult to provide formal guarantees about recommendation quality. We propose CARE, an adaptive calibration framework that wraps an arbitrary backbone recommender and outputs variable-size recommendation sets with finite-sample performance guarantees over interaction streams. CARE combines (i) a loss-based monitoring module that localizes behavioral changes and triggers threshold recalibration, and (ii) an online aggregation rule that promotes compact recommendation sets by dynamically reweighting candidate set predictors. We provide theoretical results establishing finite-sample guarantees for utility-based risk control and bounds on the expected set size relative to the best constituent predictor. Experiments across multiple datasets and backbone models demonstrate that CARE improves robustness and maintains compact recommendation sets while preserving the desired statistical guarantees. The code and implementation are available in https://github.com/kalpiree/CARE.
APA
Bisht, N., Huo, H., Gong, X. & Xu, G.. (2026). CARE: Adaptive Calibration for Reliable Recommendations. Proceedings of the 43rd International Conference on Machine Learning, in Proceedings of Machine Learning Research 306:8340-8372 Available from https://proceedings.mlr.press/v306/bisht26b.html.

Related Material