Classification Utility, Fairness, and Compactness via Tunable Information Bottleneck and R\'enyi Measures

TL;DR

This paper introduces RFIB, a novel fair representation learning method that balances utility, fairness, and compactness using tunable information bottleneck and Re9nyi measures, outperforming existing approaches on multiple datasets.

Contribution

The paper proposes RFIB, incorporating both demographic parity and equalized odds constraints, with a variational approach using Re9nyi measures to control compactness and fairness in learned representations.

Findings

RFIB outperforms state-of-the-art methods on diverse datasets.

The e9psilon parameter offers additional control over compactness.

RFIB effectively balances utility and fairness metrics.

Abstract

Designing machine learning algorithms that are accurate yet fair, not discriminating based on any sensitive attribute, is of paramount importance for society to accept AI for critical applications. In this article, we propose a novel fair representation learning method termed the R\'enyi Fair Information Bottleneck Method (RFIB) which incorporates constraints for utility, fairness, and compactness (compression) of representation, and apply it to image and tabular data classification. A key attribute of our approach is that we consider - in contrast to most prior work - both demographic parity and equalized odds as fairness constraints, allowing for a more nuanced satisfaction of both criteria. Leveraging a variational approach, we show that our objectives yield a loss function involving classical Information Bottleneck (IB) measures and establish an upper bound in terms of two R\'enyi measures of order $\alpha$ on the mutual information IB term measuring compactness between the input and its encoded embedding. We study the influence of the $\alpha$ parameter as well as two other tunable IB parameters on achieving utility/fairness trade-off goals, and show that the $\alpha$ parameter gives an additional degree of freedom that can be used to control the compactness of the representation. Experimenting on three different image datasets (EyePACS, CelebA, and FairFace) and two tabular datasets (Adult and COMPAS), using both binary and categorical sensitive attributes, we show that on various utility, fairness, and compound utility/fairness metrics RFIB outperforms current state-of-the-art approaches.