On the Impossibility of Fairness-Aware Learning from Corrupted Data

Nikola Konstantinov, Christoph H. Lampert
Proceedings of The Algorithmic Fairness through the Lens of Causality and Robustness, PMLR 171:59-83, 2022.

Abstract

Addressing fairness concerns about machine learning models is a crucial step towards their long-term adoption in real-world automated systems. Many approaches for training fair models from data have been developed and an implicit assumption about such algorithms is that they are able to recover a fair model, despite potential historical biases in the data. In this work we show a number of impossibility results that indicate that there is no learning algorithm that can recover a fair model when a proportion of the dataset is subject to arbitrary manipulations. Specifically, we prove that there are situations in which an adversary can force any learner to return a biased classifier, with or without degrading accuracy, and that the strength of this bias increases for learning problems with underrepresented protected groups in the data. Our results emphasize on the importance of studying further data corruption models of various strength and of establishing stricter data collection practices for fairness-aware learning.

Cite this Paper

BibTeX
@InProceedings{pmlr-v171-konstantinov22a, title = {On the Impossibility of Fairness-Aware Learning from Corrupted Data}, author = {Konstantinov, Nikola and Lampert, Christoph H.}, booktitle = {Proceedings of The Algorithmic Fairness through the Lens of Causality and Robustness}, pages = {59--83}, year = {2022}, editor = {Schrouff, Jessica and Dieng, Awa and Rateike, Miriam and Kwegyir-Aggrey, Kweku and Farnadi, Golnoosh}, volume = {171}, series = {Proceedings of Machine Learning Research}, month = {13 Dec}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v171/konstantinov22a/konstantinov22a.pdf}, url = {https://proceedings.mlr.press/v171/konstantinov22a.html}, abstract = {Addressing fairness concerns about machine learning models is a crucial step towards their long-term adoption in real-world automated systems. Many approaches for training fair models from data have been developed and an implicit assumption about such algorithms is that they are able to recover a fair model, despite potential historical biases in the data. In this work we show a number of impossibility results that indicate that there is no learning algorithm that can recover a fair model when a proportion of the dataset is subject to arbitrary manipulations. Specifically, we prove that there are situations in which an adversary can force any learner to return a biased classifier, with or without degrading accuracy, and that the strength of this bias increases for learning problems with underrepresented protected groups in the data. Our results emphasize on the importance of studying further data corruption models of various strength and of establishing stricter data collection practices for fairness-aware learning.} }
Endnote
%0 Conference Paper %T On the Impossibility of Fairness-Aware Learning from Corrupted Data %A Nikola Konstantinov %A Christoph H. Lampert %B Proceedings of The Algorithmic Fairness through the Lens of Causality and Robustness %C Proceedings of Machine Learning Research %D 2022 %E Jessica Schrouff %E Awa Dieng %E Miriam Rateike %E Kweku Kwegyir-Aggrey %E Golnoosh Farnadi %F pmlr-v171-konstantinov22a %I PMLR %P 59--83 %U https://proceedings.mlr.press/v171/konstantinov22a.html %V 171 %X Addressing fairness concerns about machine learning models is a crucial step towards their long-term adoption in real-world automated systems. Many approaches for training fair models from data have been developed and an implicit assumption about such algorithms is that they are able to recover a fair model, despite potential historical biases in the data. In this work we show a number of impossibility results that indicate that there is no learning algorithm that can recover a fair model when a proportion of the dataset is subject to arbitrary manipulations. Specifically, we prove that there are situations in which an adversary can force any learner to return a biased classifier, with or without degrading accuracy, and that the strength of this bias increases for learning problems with underrepresented protected groups in the data. Our results emphasize on the importance of studying further data corruption models of various strength and of establishing stricter data collection practices for fairness-aware learning.
APA
Konstantinov, N. & Lampert, C.H.. (2022). On the Impossibility of Fairness-Aware Learning from Corrupted Data. Proceedings of The Algorithmic Fairness through the Lens of Causality and Robustness, in Proceedings of Machine Learning Research 171:59-83 Available from https://proceedings.mlr.press/v171/konstantinov22a.html.

Related Material