Fair Generalized Linear Models with a Convex Penalty

Hyungrok Do, Preston Putzel, Axel S Martin, Padhraic Smyth, Judy Zhong
Proceedings of the 39th International Conference on Machine Learning, PMLR 162:5286-5308, 2022.

Abstract

Despite recent advances in algorithmic fairness, methodologies for achieving fairness with generalized linear models (GLMs) have yet to be explored in general, despite GLMs being widely used in practice. In this paper we introduce two fairness criteria for GLMs based on equalizing expected outcomes or log-likelihoods. We prove that for GLMs both criteria can be achieved via a convex penalty term based solely on the linear components of the GLM, thus permitting efficient optimization. We also derive theoretical properties for the resulting fair GLM estimator. To empirically demonstrate the efficacy of the proposed fair GLM, we compare it with other well-known fair prediction methods on an extensive set of benchmark datasets for binary classification and regression. In addition, we demonstrate that the fair GLM can generate fair predictions for a range of response variables, other than binary and continuous outcomes.

Cite this Paper

BibTeX
@InProceedings{pmlr-v162-do22a, title = {Fair Generalized Linear Models with a Convex Penalty}, author = {Do, Hyungrok and Putzel, Preston and Martin, Axel S and Smyth, Padhraic and Zhong, Judy}, booktitle = {Proceedings of the 39th International Conference on Machine Learning}, pages = {5286--5308}, year = {2022}, editor = {Chaudhuri, Kamalika and Jegelka, Stefanie and Song, Le and Szepesvari, Csaba and Niu, Gang and Sabato, Sivan}, volume = {162}, series = {Proceedings of Machine Learning Research}, month = {17--23 Jul}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v162/do22a/do22a.pdf}, url = {https://proceedings.mlr.press/v162/do22a.html}, abstract = {Despite recent advances in algorithmic fairness, methodologies for achieving fairness with generalized linear models (GLMs) have yet to be explored in general, despite GLMs being widely used in practice. In this paper we introduce two fairness criteria for GLMs based on equalizing expected outcomes or log-likelihoods. We prove that for GLMs both criteria can be achieved via a convex penalty term based solely on the linear components of the GLM, thus permitting efficient optimization. We also derive theoretical properties for the resulting fair GLM estimator. To empirically demonstrate the efficacy of the proposed fair GLM, we compare it with other well-known fair prediction methods on an extensive set of benchmark datasets for binary classification and regression. In addition, we demonstrate that the fair GLM can generate fair predictions for a range of response variables, other than binary and continuous outcomes.} }
Endnote
%0 Conference Paper %T Fair Generalized Linear Models with a Convex Penalty %A Hyungrok Do %A Preston Putzel %A Axel S Martin %A Padhraic Smyth %A Judy Zhong %B Proceedings of the 39th International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2022 %E Kamalika Chaudhuri %E Stefanie Jegelka %E Le Song %E Csaba Szepesvari %E Gang Niu %E Sivan Sabato %F pmlr-v162-do22a %I PMLR %P 5286--5308 %U https://proceedings.mlr.press/v162/do22a.html %V 162 %X Despite recent advances in algorithmic fairness, methodologies for achieving fairness with generalized linear models (GLMs) have yet to be explored in general, despite GLMs being widely used in practice. In this paper we introduce two fairness criteria for GLMs based on equalizing expected outcomes or log-likelihoods. We prove that for GLMs both criteria can be achieved via a convex penalty term based solely on the linear components of the GLM, thus permitting efficient optimization. We also derive theoretical properties for the resulting fair GLM estimator. To empirically demonstrate the efficacy of the proposed fair GLM, we compare it with other well-known fair prediction methods on an extensive set of benchmark datasets for binary classification and regression. In addition, we demonstrate that the fair GLM can generate fair predictions for a range of response variables, other than binary and continuous outcomes.
APA
Do, H., Putzel, P., Martin, A.S., Smyth, P. & Zhong, J.. (2022). Fair Generalized Linear Models with a Convex Penalty. Proceedings of the 39th International Conference on Machine Learning, in Proceedings of Machine Learning Research 162:5286-5308 Available from https://proceedings.mlr.press/v162/do22a.html.

Related Material