Fair Classification by Direct Intervention on Operating Characteristics
This work addresses fairness in machine learning for applications like risk assessment, offering a novel post-processing technique that is incremental in improving existing methods.
The paper tackles the problem of group fairness in binary classification by developing a method that directly intervenes on the operating characteristics of a pre-trained classifier to satisfy multiple fairness constraints like demographic parity, equalized odds, and predictive parity. The result shows that on datasets such as COMPAS and ACSIncome, the approach achieves approximate fairness with few interventions and minimal accuracy drop, outperforming previous methods.
We develop new classifiers under group fairness in the attribute-aware setting for binary classification with multiple group fairness constraints (e.g., demographic parity (DP), equalized odds (EO), and predictive parity (PP)). We propose a novel approach, applicable to linear fractional constraints, based on directly intervening on the operating characteristics of a pre-trained base classifier, by (i) identifying optimal operating characteristics using the base classifier's group-wise ROC convex hulls and (ii) post-processing the base classifier to match those targets. As practical post-processors, we consider randomizing a mixture of group-wise thresholding rules subject to minimizing the expected number of interventions. We further extend our approach to handle multiple protected attributes and multiple linear fractional constraints. On standard datasets (COMPAS and ACSIncome), our methods simultaneously satisfy approximate DP, EO, and PP with few interventions and a near-oracle drop in accuracy; comparing favorably to previous methods.