Information-Theoretic Bias Reduction via Causal View of Spurious Correlation

TL;DR

This paper introduces an information-theoretic approach to measure and reduce algorithmic bias by leveraging causal interpretations of spurious correlations, validated through extensive experiments.

Contribution

It presents a novel bias measurement method using conditional mutual information and a new debiasing framework incorporating this measure, including an unsupervised technique based on label noise.

Findings

Effective bias measurement via causal information theory

Improved debiasing performance on standard benchmarks

Unsupervised bias mitigation without explicit bias labels

Abstract

We propose an information-theoretic bias measurement technique through a causal interpretation of spurious correlation, which is effective to identify the feature-level algorithmic bias by taking advantage of conditional mutual information. Although several bias measurement methods have been proposed and widely investigated to achieve algorithmic fairness in various tasks such as face recognition, their accuracy- or logit-based metrics are susceptible to leading to trivial prediction score adjustment rather than fundamental bias reduction. Hence, we design a novel debiasing framework against the algorithmic bias, which incorporates a bias regularization loss derived by the proposed information-theoretic bias measurement approach. In addition, we present a simple yet effective unsupervised debiasing technique based on stochastic label noise, which does not require the explicit supervision of bias information. The proposed bias measurement and debiasing approaches are validated in diverse realistic scenarios through extensive experiments on multiple standard benchmarks.