Input-agnostic Certified Group Fairness via Gaussian Parameter Smoothing

Jiayin Jin, Zeru Zhang, Yang Zhou, Lingfei Wu
Proceedings of the 39th International Conference on Machine Learning, PMLR 162:10340-10361, 2022.

Abstract

Only recently, researchers attempt to provide classification algorithms with provable group fairness guarantees. Most of these algorithms suffer from harassment caused by the requirement that the training and deployment data follow the same distribution. This paper proposes an input-agnostic certified group fairness algorithm, FairSmooth, for improving the fairness of classification models while maintaining the remarkable prediction accuracy. A Gaussian parameter smoothing method is developed to transform base classifiers into their smooth versions. An optimal individual smooth classifier is learnt for each group with only the data regarding the group and an overall smooth classifier for all groups is generated by averaging the parameters of all the individual smooth ones. By leveraging the theory of nonlinear functional analysis, the smooth classifiers are reformulated as output functions of a Nemytskii operator. Theoretical analysis is conducted to derive that the Nemytskii operator is smooth and induces a Frechet differentiable smooth manifold. We theoretically demonstrate that the smooth manifold has a global Lipschitz constant that is independent of the domain of the input data, which derives the input-agnostic certified group fairness.

Cite this Paper

BibTeX
@InProceedings{pmlr-v162-jin22g, title = {Input-agnostic Certified Group Fairness via {G}aussian Parameter Smoothing}, author = {Jin, Jiayin and Zhang, Zeru and Zhou, Yang and Wu, Lingfei}, booktitle = {Proceedings of the 39th International Conference on Machine Learning}, pages = {10340--10361}, 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/jin22g/jin22g.pdf}, url = {https://proceedings.mlr.press/v162/jin22g.html}, abstract = {Only recently, researchers attempt to provide classification algorithms with provable group fairness guarantees. Most of these algorithms suffer from harassment caused by the requirement that the training and deployment data follow the same distribution. This paper proposes an input-agnostic certified group fairness algorithm, FairSmooth, for improving the fairness of classification models while maintaining the remarkable prediction accuracy. A Gaussian parameter smoothing method is developed to transform base classifiers into their smooth versions. An optimal individual smooth classifier is learnt for each group with only the data regarding the group and an overall smooth classifier for all groups is generated by averaging the parameters of all the individual smooth ones. By leveraging the theory of nonlinear functional analysis, the smooth classifiers are reformulated as output functions of a Nemytskii operator. Theoretical analysis is conducted to derive that the Nemytskii operator is smooth and induces a Frechet differentiable smooth manifold. We theoretically demonstrate that the smooth manifold has a global Lipschitz constant that is independent of the domain of the input data, which derives the input-agnostic certified group fairness.} }
Endnote
%0 Conference Paper %T Input-agnostic Certified Group Fairness via Gaussian Parameter Smoothing %A Jiayin Jin %A Zeru Zhang %A Yang Zhou %A Lingfei Wu %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-jin22g %I PMLR %P 10340--10361 %U https://proceedings.mlr.press/v162/jin22g.html %V 162 %X Only recently, researchers attempt to provide classification algorithms with provable group fairness guarantees. Most of these algorithms suffer from harassment caused by the requirement that the training and deployment data follow the same distribution. This paper proposes an input-agnostic certified group fairness algorithm, FairSmooth, for improving the fairness of classification models while maintaining the remarkable prediction accuracy. A Gaussian parameter smoothing method is developed to transform base classifiers into their smooth versions. An optimal individual smooth classifier is learnt for each group with only the data regarding the group and an overall smooth classifier for all groups is generated by averaging the parameters of all the individual smooth ones. By leveraging the theory of nonlinear functional analysis, the smooth classifiers are reformulated as output functions of a Nemytskii operator. Theoretical analysis is conducted to derive that the Nemytskii operator is smooth and induces a Frechet differentiable smooth manifold. We theoretically demonstrate that the smooth manifold has a global Lipschitz constant that is independent of the domain of the input data, which derives the input-agnostic certified group fairness.
APA
Jin, J., Zhang, Z., Zhou, Y. & Wu, L.. (2022). Input-agnostic Certified Group Fairness via Gaussian Parameter Smoothing. Proceedings of the 39th International Conference on Machine Learning, in Proceedings of Machine Learning Research 162:10340-10361 Available from https://proceedings.mlr.press/v162/jin22g.html.

Related Material