Information-Theoretic Bias Reduction via Causal View of Spurious Correlation
This addresses algorithmic fairness issues in machine learning applications, offering a method to reduce bias without explicit supervision, though it appears incremental relative to existing bias measurement approaches.
The paper tackles algorithmic bias in tasks like face recognition by proposing an information-theoretic bias measurement technique based on causal interpretation of spurious correlation, and introduces a debiasing framework with regularization loss and unsupervised technique, validated on multiple benchmarks.
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.