Feature Importance Disparities for Data Bias Investigations
This addresses the need for automated tools to assist practitioners in investigating data bias, which is crucial for fairness in machine learning applications.
The paper tackles the problem of identifying data bias in classifiers by introducing a method to detect subgroups where feature importance significantly differs from the overall dataset, showing across 4 datasets and 4 feature importance methods that these subgroups often correspond to serious bias issues as measured by fairness metrics.
It is widely held that one cause of downstream bias in classifiers is bias present in the training data. Rectifying such biases may involve context-dependent interventions such as training separate models on subgroups, removing features with bias in the collection process, or even conducting real-world experiments to ascertain sources of bias. Despite the need for such data bias investigations, few automated methods exist to assist practitioners in these efforts. In this paper, we present one such method that given a dataset $X$ consisting of protected and unprotected features, outcomes $y$, and a regressor $h$ that predicts $y$ given $X$, outputs a tuple $(f_j, g)$, with the following property: $g$ corresponds to a subset of the training dataset $(X, y)$, such that the $j^{th}$ feature $f_j$ has much larger (or smaller) influence in the subgroup $g$, than on the dataset overall, which we call feature importance disparity (FID). We show across $4$ datasets and $4$ common feature importance methods of broad interest to the machine learning community that we can efficiently find subgroups with large FID values even over exponentially large subgroup classes and in practice these groups correspond to subgroups with potentially serious bias issues as measured by standard fairness metrics.