Renyi Fair Information Bottleneck for Image Classification

TL;DR

This paper introduces the Renyi Fair Information Bottleneck (RFIB), a novel method that uses Renyi divergence to learn fair and compact representations for image classification, balancing utility and fairness constraints.

Contribution

The paper proposes a new variational loss function based on Renyi divergence for fairness constraints, applicable to image classification, and demonstrates superior performance over existing methods.

Findings

Outperforms state-of-the-art fairness techniques on EyePACS dataset

Balances utility and fairness effectively with Renyi divergence

Achieves lower accuracy gap and higher minimal accuracy

Abstract

We develop a novel method for ensuring fairness in machine learning which we term as the Renyi Fair Information Bottleneck (RFIB). We consider two different fairness constraints - demographic parity and equalized odds - for learning fair representations and derive a loss function via a variational approach that uses Renyi's divergence with its tunable parameter $\alpha$ and that takes into account the triple constraints of utility, fairness, and compactness of representation. We then evaluate the performance of our method for image classification using the EyePACS medical imaging dataset, showing it outperforms competing state of the art techniques with performance measured using a variety of compound utility/fairness metrics, including accuracy gap and Rawls' minimal accuracy.