A post-processing algorithm achieves distribution-free finite-sample group fairness guarantees with controlled excess risk for both group-aware and group-blind settings, shown minimax-optimal up to logs via lower bound.
(23) Therefore, when D0 ≤ −˜ϵG η − 2ϵα, we have ˜α = α and λ∗ ˜α = λ∗ α = 0
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
stat.ME 1years
2024 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Finite-Sample and Distribution-Free Fair Classification: Optimal Trade-off Between Excess Risk and Fairness, and the Cost of Group-Blindness
A post-processing algorithm achieves distribution-free finite-sample group fairness guarantees with controlled excess risk for both group-aware and group-blind settings, shown minimax-optimal up to logs via lower bound.