Fair Neighbor Embedding

Jaakko Peltonen, Wen Xu, Timo Nummenmaa, Jyrki Nummenmaa
Proceedings of the 40th International Conference on Machine Learning, PMLR 202:27564-27584, 2023.

Abstract

We consider fairness in dimensionality reduction. Nonlinear dimensionality reduction yields low dimensional representations that let users visualize and explore high-dimensional data. However, traditional dimensionality reduction may yield biased visualizations overemphasizing relationships of societal phenomena to sensitive attributes or protected groups. We introduce a framework of fair neighbor embedding, the Fair Neighbor Retrieval Visualizer, which formulates fair nonlinear dimensionality reduction as an information retrieval task whose performance and fairness are quantified by information retrieval criteria. The method optimizes low-dimensional embeddings that preserve high-dimensional data neighborhoods without yielding biased association of such neighborhoods to protected groups. In experiments the method yields fair visualizations outperforming previous methods.

Cite this Paper

BibTeX
@InProceedings{pmlr-v202-peltonen23a, title = {Fair Neighbor Embedding}, author = {Peltonen, Jaakko and Xu, Wen and Nummenmaa, Timo and Nummenmaa, Jyrki}, booktitle = {Proceedings of the 40th International Conference on Machine Learning}, pages = {27564--27584}, year = {2023}, editor = {Krause, Andreas and Brunskill, Emma and Cho, Kyunghyun and Engelhardt, Barbara and Sabato, Sivan and Scarlett, Jonathan}, volume = {202}, series = {Proceedings of Machine Learning Research}, month = {23--29 Jul}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v202/peltonen23a/peltonen23a.pdf}, url = {https://proceedings.mlr.press/v202/peltonen23a.html}, abstract = {We consider fairness in dimensionality reduction. Nonlinear dimensionality reduction yields low dimensional representations that let users visualize and explore high-dimensional data. However, traditional dimensionality reduction may yield biased visualizations overemphasizing relationships of societal phenomena to sensitive attributes or protected groups. We introduce a framework of fair neighbor embedding, the Fair Neighbor Retrieval Visualizer, which formulates fair nonlinear dimensionality reduction as an information retrieval task whose performance and fairness are quantified by information retrieval criteria. The method optimizes low-dimensional embeddings that preserve high-dimensional data neighborhoods without yielding biased association of such neighborhoods to protected groups. In experiments the method yields fair visualizations outperforming previous methods.} }
Endnote
%0 Conference Paper %T Fair Neighbor Embedding %A Jaakko Peltonen %A Wen Xu %A Timo Nummenmaa %A Jyrki Nummenmaa %B Proceedings of the 40th International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2023 %E Andreas Krause %E Emma Brunskill %E Kyunghyun Cho %E Barbara Engelhardt %E Sivan Sabato %E Jonathan Scarlett %F pmlr-v202-peltonen23a %I PMLR %P 27564--27584 %U https://proceedings.mlr.press/v202/peltonen23a.html %V 202 %X We consider fairness in dimensionality reduction. Nonlinear dimensionality reduction yields low dimensional representations that let users visualize and explore high-dimensional data. However, traditional dimensionality reduction may yield biased visualizations overemphasizing relationships of societal phenomena to sensitive attributes or protected groups. We introduce a framework of fair neighbor embedding, the Fair Neighbor Retrieval Visualizer, which formulates fair nonlinear dimensionality reduction as an information retrieval task whose performance and fairness are quantified by information retrieval criteria. The method optimizes low-dimensional embeddings that preserve high-dimensional data neighborhoods without yielding biased association of such neighborhoods to protected groups. In experiments the method yields fair visualizations outperforming previous methods.
APA
Peltonen, J., Xu, W., Nummenmaa, T. & Nummenmaa, J.. (2023). Fair Neighbor Embedding. Proceedings of the 40th International Conference on Machine Learning, in Proceedings of Machine Learning Research 202:27564-27584 Available from https://proceedings.mlr.press/v202/peltonen23a.html.

Related Material