Conformal Classification with Equalized Coverage for Adaptively Selected Groups

TL;DR

This paper proposes a conformal inference approach that provides valid, group-sensitive prediction sets to balance accuracy and fairness in classification tasks.

Contribution

It introduces a method for conformal classification that ensures equalized coverage across adaptively selected groups, addressing fairness and efficiency simultaneously.

Findings

Method achieves valid coverage conditional on selected features.

Demonstrates effectiveness on simulated data.

Ensures fairness by equalizing coverage for sensitive groups.

Abstract

This paper introduces a conformal inference method to evaluate uncertainty in classification by generating prediction sets with valid coverage conditional on adaptively chosen features. These features are carefully selected to reflect potential model limitations or biases. This can be useful to find a practical compromise between efficiency -- by providing informative predictions -- and algorithmic fairness -- by ensuring equalized coverage for the most sensitive groups. We demonstrate the validity and effectiveness of this method on simulated and real data sets.