TY - GEN
T1 - A Reduction to Binary Approach for Debiasing Multiclass Datasets
AU - Alabdulmohsin, Ibrahim
AU - Schrouff, Jessica
AU - Koyejo, Oluwasanmi
N1 - The authors would like to acknowledge and thank Yuan Liu and Alex Brown at Google Health AI for their support with the dermatology application, and thank Kathy Meier-Hellstern and Lucas Dixon from Google Responsible AI team for their feedback on earlier drafts of this manuscript.
PY - 2022
Y1 - 2022
N2 - We propose a novel reduction-to-binary (R2B) approach that enforces demographic parity for multiclass classification with non-binary sensitive attributes via a reduction to a sequence of binary debiasing tasks. We prove that R2B satisfies optimality and bias guarantees and demonstrate empirically that it can lead to an improvement over two baselines: (1) treating multiclass problems as multi-label by debiasing labels independently and (2) transforming the features instead of the labels. Surprisingly, we also demonstrate that independent label debiasing yields competitive results in most (but not all) settings. We validate these conclusions on synthetic and real-world datasets from social science, computer vision, and healthcare.
AB - We propose a novel reduction-to-binary (R2B) approach that enforces demographic parity for multiclass classification with non-binary sensitive attributes via a reduction to a sequence of binary debiasing tasks. We prove that R2B satisfies optimality and bias guarantees and demonstrate empirically that it can lead to an improvement over two baselines: (1) treating multiclass problems as multi-label by debiasing labels independently and (2) transforming the features instead of the labels. Surprisingly, we also demonstrate that independent label debiasing yields competitive results in most (but not all) settings. We validate these conclusions on synthetic and real-world datasets from social science, computer vision, and healthcare.
UR - http://www.scopus.com/inward/record.url?scp=85163207056&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85163207056&partnerID=8YFLogxK
M3 - Conference contribution
AN - SCOPUS:85163207056
T3 - Advances in Neural Information Processing Systems
BT - Advances in Neural Information Processing Systems 35 - 36th Conference on Neural Information Processing Systems, NeurIPS 2022
A2 - Koyejo, S.
A2 - Mohamed, S.
A2 - Agarwal, A.
A2 - Belgrave, D.
A2 - Cho, K.
A2 - Oh, A.
PB - Neural information processing systems foundation
T2 - 36th Conference on Neural Information Processing Systems, NeurIPS 2022
Y2 - 28 November 2022 through 9 December 2022
ER -