Fair Decisions Despite Imperfect Predictions

Niki Kilbertus, Manuel Gomez Rodriguez, Bernhard Schölkopf, Krikamol Muandet, Isabel Valera
Proceedings of the Twenty Third International Conference on Artificial Intelligence and Statistics, PMLR 108:277-287, 2020.

Abstract

Consequential decisions are increasingly informed by sophisticated data-driven predictive models. However, consistently learning accurate predictive models requires access to ground truth labels. Unfortunately, in practice, labels may only exist conditional on certain decisions—if a loan is denied, there is not even an option for the individual to pay back the loan. In this paper, we show that, in this selective labels setting, learning to predict is suboptimal in terms of both fairness and utility. To avoid this undesirable behavior, we propose to directly learn stochastic decision policies that maximize utility under fairness constraints. In the context of fair machine learning, our results suggest the need for a paradigm shift from "learning to predict" to "learning to decide". Experiments on synthetic and real-world data illustrate the favorable properties of learning to decide, in terms of both utility and fairness.

Cite this Paper

BibTeX
@InProceedings{pmlr-v108-kilbertus20a, title = {Fair Decisions Despite Imperfect Predictions}, author = {Kilbertus, Niki and Rodriguez, Manuel Gomez and Sch\"olkopf, Bernhard and Muandet, Krikamol and Valera, Isabel}, booktitle = {Proceedings of the Twenty Third International Conference on Artificial Intelligence and Statistics}, pages = {277--287}, year = {2020}, editor = {Chiappa, Silvia and Calandra, Roberto}, volume = {108}, series = {Proceedings of Machine Learning Research}, month = {26--28 Aug}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v108/kilbertus20a/kilbertus20a.pdf}, url = {https://proceedings.mlr.press/v108/kilbertus20a.html}, abstract = {Consequential decisions are increasingly informed by sophisticated data-driven predictive models. However, consistently learning accurate predictive models requires access to ground truth labels. Unfortunately, in practice, labels may only exist conditional on certain decisions—if a loan is denied, there is not even an option for the individual to pay back the loan. In this paper, we show that, in this selective labels setting, learning to predict is suboptimal in terms of both fairness and utility. To avoid this undesirable behavior, we propose to directly learn stochastic decision policies that maximize utility under fairness constraints. In the context of fair machine learning, our results suggest the need for a paradigm shift from "learning to predict" to "learning to decide". Experiments on synthetic and real-world data illustrate the favorable properties of learning to decide, in terms of both utility and fairness.} }
Endnote
%0 Conference Paper %T Fair Decisions Despite Imperfect Predictions %A Niki Kilbertus %A Manuel Gomez Rodriguez %A Bernhard Schölkopf %A Krikamol Muandet %A Isabel Valera %B Proceedings of the Twenty Third International Conference on Artificial Intelligence and Statistics %C Proceedings of Machine Learning Research %D 2020 %E Silvia Chiappa %E Roberto Calandra %F pmlr-v108-kilbertus20a %I PMLR %P 277--287 %U https://proceedings.mlr.press/v108/kilbertus20a.html %V 108 %X Consequential decisions are increasingly informed by sophisticated data-driven predictive models. However, consistently learning accurate predictive models requires access to ground truth labels. Unfortunately, in practice, labels may only exist conditional on certain decisions—if a loan is denied, there is not even an option for the individual to pay back the loan. In this paper, we show that, in this selective labels setting, learning to predict is suboptimal in terms of both fairness and utility. To avoid this undesirable behavior, we propose to directly learn stochastic decision policies that maximize utility under fairness constraints. In the context of fair machine learning, our results suggest the need for a paradigm shift from "learning to predict" to "learning to decide". Experiments on synthetic and real-world data illustrate the favorable properties of learning to decide, in terms of both utility and fairness.
APA
Kilbertus, N., Rodriguez, M.G., Schölkopf, B., Muandet, K. & Valera, I.. (2020). Fair Decisions Despite Imperfect Predictions. Proceedings of the Twenty Third International Conference on Artificial Intelligence and Statistics, in Proceedings of Machine Learning Research 108:277-287 Available from https://proceedings.mlr.press/v108/kilbertus20a.html.

Related Material