Multi-Group Fairness Evaluation via Conditional Value-at-Risk Testing

TL;DR

This paper introduces a CVaR-based testing method to evaluate ML model fairness across multiple sensitive groups, significantly reducing sample complexity and addressing challenges of high-dimensional group attributes.

Contribution

It proposes a novel CVaR testing approach for fairness evaluation that lowers sample complexity exponentially and incorporates prior distribution and non-i.i.d. strategies.

Findings

Sample complexity is reduced exponentially with CVaR testing.

Re9nyi entropy characterizes the sample complexity under prior weighting.

Non-i.i.d. data collection can make sample complexity independent of group count.

Abstract

Machine learning (ML) models used in prediction and classification tasks may display performance disparities across population groups determined by sensitive attributes (e.g., race, sex, age). We consider the problem of evaluating the performance of a fixed ML model across population groups defined by multiple sensitive attributes (e.g., race and sex and age). Here, the sample complexity for estimating the worst-case performance gap across groups (e.g., the largest difference in error rates) increases exponentially with the number of group-denoting sensitive attributes. To address this issue, we propose an approach to test for performance disparities based on Conditional Value-at-Risk (CVaR). By allowing a small probabilistic slack on the groups over which a model has approximately equal performance, we show that the sample complexity required for discovering performance violations is reduced exponentially to be at most upper bounded by the square root of the number of groups. As a byproduct of our analysis, when the groups are weighted by a specific prior distribution, we show that R\'enyi entropy of order 2/3 of the prior distribution captures the sample complexity of the proposed CVaR test algorithm. Finally, we also show that there exists a non-i.i.d. data collection strategy that results in a sample complexity independent of the number of groups.