PLS-based approach for fair representation learning

2025-02-22

AI Summary

This paper addresses the challenge of jointly optimizing dimensionality reduction and fairness in fair representation learning. We propose Fair Partial Least Squares (Fair PLS), the first method to incorporate group-level fairness constraints—such as statistical parity and equalized odds—directly into the PLS framework, enabling joint optimization of predictive performance and fairness during projection. Fair PLS supports both linear and kernel-based nonlinear extensions. Unlike fair PCA—which relies on restrictive orthogonality assumptions—Fair PLS offers greater representational capacity and a more flexible fairness–accuracy trade-off. Extensive experiments on multiple benchmark datasets demonstrate that Fair PLS not only preserves or even improves prediction accuracy but also significantly enhances various group fairness metrics, validating its effectiveness and practical utility.

Technology Category

Machine Learning: Dimensionality Reduction/Feature SelectionComputer Vision: Representation Learning for VisionNatural Language Processing: Ethics — Bias, Fairness, Transparency & Privacy

Application Category

User Modeling, Personalization and Recommendation: Fairness-aware retrieval and rankingSearch and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for rankingSocial Networks and Social Media: Fairness and bias in social network and social media analysis

Abstract

We revisit the problem of fair representation learning by proposing Fair Partial Least Squares (PLS) components. PLS is widely used in statistics to efficiently reduce the dimension of the data by providing representation tailored for the prediction. We propose a novel method to incorporate fairness constraints in the construction of PLS components. This new algorithm provides a feasible way to construct such features both in the linear and the non linear case using kernel embeddings. The efficiency of our method is evaluated on different datasets, and we prove its superiority with respect to standard fair PCA method.

Problem

• Fair representation learning
• Dimension reduction
• Fairness constraints in PLS

Innovation

• Fair Partial Least Squares
• Fairness constraints integration
• Kernel embeddings utilization