Counterfactually Fair Conformal Prediction

Ozgur Guldogan, Neeraj Sarna, Yuanyuan Li, Michael Berger
Proceedings of The 29th International Conference on Artificial Intelligence and Statistics, PMLR 300:4717-4725, 2026.

Abstract

While counterfactual fairness of point predictors is well studied, its extension to prediction \emph{sets}—central to fair decision-making under uncertainty—remains underexplored. On the other hand, conformal prediction (CP) provides efficient, distribution-free, finite-sample valid prediction sets, yet does not ensure counterfactual fairness. We close this gap by developing \emph{Counterfactually Fair Conformal Prediction} (CF-CP) that produces counterfactually fair prediction sets. Through symmetrization of conformity scores across protected-attribute interventions, we prove that CF-CP results in counterfactually fair prediction sets while maintaining the marginal coverage property. Furthermore, we empirically demonstrate that on both synthetic and real datasets, across regression and classification tasks, CF-CP achieves the desired counterfactual fairness and meets the target coverage rate with minimal increase in prediction set size. CF-CP offers a simple, training-free route to counterfactually fair uncertainty quantification.

Cite this Paper


BibTeX
@InProceedings{pmlr-v300-guldogan26a, title = { Counterfactually Fair Conformal Prediction }, author = {Guldogan, Ozgur and Sarna, Neeraj and Li, Yuanyuan and Berger, Michael}, booktitle = {Proceedings of The 29th International Conference on Artificial Intelligence and Statistics}, pages = {4717--4725}, year = {2026}, editor = {Khan, Emtiyaz and Li, Yingzhen and Solin, Arno and Ramdas, Aaditya}, volume = {300}, series = {Proceedings of Machine Learning Research}, month = {02--05 May}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v300/main/assets/guldogan26a/guldogan26a.pdf}, url = {https://proceedings.mlr.press/v300/guldogan26a.html}, abstract = { While counterfactual fairness of point predictors is well studied, its extension to prediction \emph{sets}—central to fair decision-making under uncertainty—remains underexplored. On the other hand, conformal prediction (CP) provides efficient, distribution-free, finite-sample valid prediction sets, yet does not ensure counterfactual fairness. We close this gap by developing \emph{Counterfactually Fair Conformal Prediction} (CF-CP) that produces counterfactually fair prediction sets. Through symmetrization of conformity scores across protected-attribute interventions, we prove that CF-CP results in counterfactually fair prediction sets while maintaining the marginal coverage property. Furthermore, we empirically demonstrate that on both synthetic and real datasets, across regression and classification tasks, CF-CP achieves the desired counterfactual fairness and meets the target coverage rate with minimal increase in prediction set size. CF-CP offers a simple, training-free route to counterfactually fair uncertainty quantification. } }
Endnote
%0 Conference Paper %T Counterfactually Fair Conformal Prediction %A Ozgur Guldogan %A Neeraj Sarna %A Yuanyuan Li %A Michael Berger %B Proceedings of The 29th International Conference on Artificial Intelligence and Statistics %C Proceedings of Machine Learning Research %D 2026 %E Emtiyaz Khan %E Yingzhen Li %E Arno Solin %E Aaditya Ramdas %F pmlr-v300-guldogan26a %I PMLR %P 4717--4725 %U https://proceedings.mlr.press/v300/guldogan26a.html %V 300 %X While counterfactual fairness of point predictors is well studied, its extension to prediction \emph{sets}—central to fair decision-making under uncertainty—remains underexplored. On the other hand, conformal prediction (CP) provides efficient, distribution-free, finite-sample valid prediction sets, yet does not ensure counterfactual fairness. We close this gap by developing \emph{Counterfactually Fair Conformal Prediction} (CF-CP) that produces counterfactually fair prediction sets. Through symmetrization of conformity scores across protected-attribute interventions, we prove that CF-CP results in counterfactually fair prediction sets while maintaining the marginal coverage property. Furthermore, we empirically demonstrate that on both synthetic and real datasets, across regression and classification tasks, CF-CP achieves the desired counterfactual fairness and meets the target coverage rate with minimal increase in prediction set size. CF-CP offers a simple, training-free route to counterfactually fair uncertainty quantification.
APA
Guldogan, O., Sarna, N., Li, Y. & Berger, M.. (2026). Counterfactually Fair Conformal Prediction . Proceedings of The 29th International Conference on Artificial Intelligence and Statistics, in Proceedings of Machine Learning Research 300:4717-4725 Available from https://proceedings.mlr.press/v300/guldogan26a.html.

Related Material