Group-blind optimal transport to group parity and its constrained variants
This addresses fairness in ML for groups defined by protected attributes like gender or race, offering a novel approach that avoids reliance on individual attribute data, though it is incremental in the context of existing fair learning methods.
The paper tackles the problem of achieving demographic group parity in machine learning without needing protected attribute values for individual samples, by designing a group-blind projection map that aligns feature distributions of privileged and unprivileged groups, and presents numerical results on synthetic and real data.
Fairness holds a pivotal role in the realm of machine learning, particularly when it comes to addressing groups categorised by protected attributes, e.g., gender, race. Prevailing algorithms in fair learning predominantly hinge on accessibility or estimations of these protected attributes, at least in the training process. We design a single group-blind projection map that aligns the feature distributions of both groups in the source data, achieving (demographic) group parity, without requiring values of the protected attribute for individual samples in the computation of the map, as well as its use. Instead, our approach utilises the feature distributions of the privileged and unprivileged groups in a boarder population and the essential assumption that the source data are unbiased representation of the population. We present numerical results on synthetic data and real data.