Multiple Attribute Fairness: Application to Fraud Detection
This addresses fairness issues in fraud detection systems, particularly for multiple sensitive attributes, though it appears incremental as an extension of existing fairness regimes.
The authors tackled fairness in classification by proposing a new fairness measure that relaxes equal odds conditions and developing a model-agnostic heuristic to calibrate outcomes across sensitive attributes. They demonstrated its effectiveness in fraud detection, achieving fairness across multiple attribute values and comparable performance to existing two-group methods on public datasets.
We propose a fairness measure relaxing the equality conditions in the popular equal odds fairness regime for classification. We design an iterative, model-agnostic, grid-based heuristic that calibrates the outcomes per sensitive attribute value to conform to the measure. The heuristic is designed to handle high arity attribute values and performs a per attribute sanitization of outcomes across different protected attribute values. We also extend our heuristic for multiple attributes. Highlighting our motivating application, fraud detection, we show that the proposed heuristic is able to achieve fairness across multiple values of a single protected attribute, multiple protected attributes. When compared to current fairness techniques, that focus on two groups, we achieve comparable performance across several public data sets.