A distributionally robust AUC-fairness method with TV-distance bounds that preserves group fairness when protected-group labels are noisy.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Preserving AUC Fairness in Learning with Noisy Protected Groups
A distributionally robust AUC-fairness method with TV-distance bounds that preserves group fairness when protected-group labels are noisy.