DeBayes: a Bayesian Method for Debiasing Network Embeddings

Maarten Buyl, Tijl De Bie
Proceedings of the 37th International Conference on Machine Learning, PMLR 119:1220-1229, 2020.

Abstract

As machine learning algorithms are increasingly deployed for high-impact automated decision making, ethical and increasingly also legal standards demand that they treat all individuals fairly, without discrimination based on their age, gender, race or other sensitive traits. In recent years much progress has been made on ensuring fairness and reducing bias in standard machine learning settings. Yet, for network embedding, with applications in vulnerable domains ranging from social network analysis to recommender systems, current options remain limited both in number and performance. We thus propose DeBayes: a conceptually elegant Bayesian method that is capable of learning debiased embeddings by using a biased prior. Our experiments show that these representations can then be used to perform link prediction that is significantly more fair in terms of popular metrics such as demographic parity and equalized opportunity.

Cite this Paper

BibTeX
@InProceedings{pmlr-v119-buyl20a, title = {{D}e{B}ayes: a {B}ayesian Method for Debiasing Network Embeddings}, author = {Buyl, Maarten and De Bie, Tijl}, booktitle = {Proceedings of the 37th International Conference on Machine Learning}, pages = {1220--1229}, year = {2020}, editor = {III, Hal Daumé and Singh, Aarti}, volume = {119}, series = {Proceedings of Machine Learning Research}, month = {13--18 Jul}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v119/buyl20a/buyl20a.pdf}, url = {https://proceedings.mlr.press/v119/buyl20a.html}, abstract = {As machine learning algorithms are increasingly deployed for high-impact automated decision making, ethical and increasingly also legal standards demand that they treat all individuals fairly, without discrimination based on their age, gender, race or other sensitive traits. In recent years much progress has been made on ensuring fairness and reducing bias in standard machine learning settings. Yet, for network embedding, with applications in vulnerable domains ranging from social network analysis to recommender systems, current options remain limited both in number and performance. We thus propose DeBayes: a conceptually elegant Bayesian method that is capable of learning debiased embeddings by using a biased prior. Our experiments show that these representations can then be used to perform link prediction that is significantly more fair in terms of popular metrics such as demographic parity and equalized opportunity.} }
Endnote
%0 Conference Paper %T DeBayes: a Bayesian Method for Debiasing Network Embeddings %A Maarten Buyl %A Tijl De Bie %B Proceedings of the 37th International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2020 %E Hal Daumé III %E Aarti Singh %F pmlr-v119-buyl20a %I PMLR %P 1220--1229 %U https://proceedings.mlr.press/v119/buyl20a.html %V 119 %X As machine learning algorithms are increasingly deployed for high-impact automated decision making, ethical and increasingly also legal standards demand that they treat all individuals fairly, without discrimination based on their age, gender, race or other sensitive traits. In recent years much progress has been made on ensuring fairness and reducing bias in standard machine learning settings. Yet, for network embedding, with applications in vulnerable domains ranging from social network analysis to recommender systems, current options remain limited both in number and performance. We thus propose DeBayes: a conceptually elegant Bayesian method that is capable of learning debiased embeddings by using a biased prior. Our experiments show that these representations can then be used to perform link prediction that is significantly more fair in terms of popular metrics such as demographic parity and equalized opportunity.
APA
Buyl, M. & De Bie, T.. (2020). DeBayes: a Bayesian Method for Debiasing Network Embeddings. Proceedings of the 37th International Conference on Machine Learning, in Proceedings of Machine Learning Research 119:1220-1229 Available from https://proceedings.mlr.press/v119/buyl20a.html.

Related Material