When do Minimax-fair Learning and Empirical Risk Minimization Coincide?

Harvineet Singh, Matthäus Kleindessner, Volkan Cevher, Rumi Chunara, Chris Russell
Proceedings of the 40th International Conference on Machine Learning, PMLR 202:31969-31989, 2023.

Abstract

Minimax-fair machine learning minimizes the error for the worst-off group. However, empirical evidence suggests that when sophisticated models are trained with standard empirical risk minimization (ERM), they often have the same performance on the worst-off group as a minimax-trained model. Our work makes this counter-intuitive observation concrete. We prove that if the hypothesis class is sufficiently expressive and the group information is recoverable from the features, ERM and minimax-fairness learning formulations indeed have the same performance on the worst-off group. We provide additional empirical evidence of how this observation holds on a wide range of datasets and hypothesis classes. Since ERM is fundamentally easier than minimax optimization, our findings have implications on the practice of fair machine learning.

Cite this Paper

BibTeX
@InProceedings{pmlr-v202-singh23b, title = {When do Minimax-fair Learning and Empirical Risk Minimization Coincide?}, author = {Singh, Harvineet and Kleindessner, Matth\"{a}us and Cevher, Volkan and Chunara, Rumi and Russell, Chris}, booktitle = {Proceedings of the 40th International Conference on Machine Learning}, pages = {31969--31989}, year = {2023}, editor = {Krause, Andreas and Brunskill, Emma and Cho, Kyunghyun and Engelhardt, Barbara and Sabato, Sivan and Scarlett, Jonathan}, volume = {202}, series = {Proceedings of Machine Learning Research}, month = {23--29 Jul}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v202/singh23b/singh23b.pdf}, url = {https://proceedings.mlr.press/v202/singh23b.html}, abstract = {Minimax-fair machine learning minimizes the error for the worst-off group. However, empirical evidence suggests that when sophisticated models are trained with standard empirical risk minimization (ERM), they often have the same performance on the worst-off group as a minimax-trained model. Our work makes this counter-intuitive observation concrete. We prove that if the hypothesis class is sufficiently expressive and the group information is recoverable from the features, ERM and minimax-fairness learning formulations indeed have the same performance on the worst-off group. We provide additional empirical evidence of how this observation holds on a wide range of datasets and hypothesis classes. Since ERM is fundamentally easier than minimax optimization, our findings have implications on the practice of fair machine learning.} }
Endnote
%0 Conference Paper %T When do Minimax-fair Learning and Empirical Risk Minimization Coincide? %A Harvineet Singh %A Matthäus Kleindessner %A Volkan Cevher %A Rumi Chunara %A Chris Russell %B Proceedings of the 40th International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2023 %E Andreas Krause %E Emma Brunskill %E Kyunghyun Cho %E Barbara Engelhardt %E Sivan Sabato %E Jonathan Scarlett %F pmlr-v202-singh23b %I PMLR %P 31969--31989 %U https://proceedings.mlr.press/v202/singh23b.html %V 202 %X Minimax-fair machine learning minimizes the error for the worst-off group. However, empirical evidence suggests that when sophisticated models are trained with standard empirical risk minimization (ERM), they often have the same performance on the worst-off group as a minimax-trained model. Our work makes this counter-intuitive observation concrete. We prove that if the hypothesis class is sufficiently expressive and the group information is recoverable from the features, ERM and minimax-fairness learning formulations indeed have the same performance on the worst-off group. We provide additional empirical evidence of how this observation holds on a wide range of datasets and hypothesis classes. Since ERM is fundamentally easier than minimax optimization, our findings have implications on the practice of fair machine learning.
APA
Singh, H., Kleindessner, M., Cevher, V., Chunara, R. & Russell, C.. (2023). When do Minimax-fair Learning and Empirical Risk Minimization Coincide?. Proceedings of the 40th International Conference on Machine Learning, in Proceedings of Machine Learning Research 202:31969-31989 Available from https://proceedings.mlr.press/v202/singh23b.html.

Related Material