Fair Data Pre-Processing with Imperfect Attribute Space

TL;DR

LatentPre is a new framework for fair data pre-processing that uses latent attributes to improve bias mitigation in imperfect attribute spaces, ensuring fairness without sacrificing utility.

Contribution

It introduces latent attributes and an EM-based estimation to enable robust fairness-aware data processing in real-world scenarios with incomplete or imperfect attribute information.

Findings

LatentPre achieves better fairness-utility trade-offs across diverse datasets.

The framework effectively captures subtle signals via latent attributes.

Extensive experiments validate its robustness and practical effectiveness.

Abstract

Fair data pre-processing is a widely used strategy for mitigating bias in machine learning. A promising line of research focuses on calibrating datasets to satisfy a designed fairness policy so that sensitive attributes influence outcomes only through clearly specified legitimate causal pathways. While effective on clean and information-rich data, these methods often break down in real-world scenarios with imperfect attribute spaces, where decision-relevant factors may be deemed unusable or even missing. To address this gap, we propose LatentPre, a novel framework that enables principled and robust fair data processing in practical settings. Instead of relying solely on observed attributes, LatentPre augments the fairness policy with latent attributes that capture essential but subtle signals, enabling the framework to operate as if the attribute space were perfect. These latent attributes are strategically introduced to guarantee identifiability and are estimated using a tailored expectation-maximization paradigm. The raw data is then carefully refined to conform to this latent-augmented policy, effectively removing biased patterns while preserving justifiable ones. Extensive experiments demonstrate that LatentPre consistently achieves strong fairness-utility trade-offs across diverse scenarios, advancing practical fairness-aware data management.