FairNet: Dynamic Fairness Correction without Performance Loss via Contrastive Conditional LoRA
This addresses fairness issues in ML models for applications where bias can harm minority groups, offering a flexible solution that is incremental in improving upon existing debiasing methods.
The paper tackles the problem of fairness in machine learning models by proposing FairNet, a framework for dynamic, instance-level fairness correction that integrates a bias detector with conditional LoRA and a contrastive loss, achieving enhanced performance for the worst group without overall performance loss as validated across vision and language benchmarks.
Ensuring fairness in machine learning models is a critical challenge. Existing debiasing methods often compromise performance, rely on static correction strategies, and struggle with data sparsity, particularly within minority groups. Furthermore, their utilization of sensitive attributes is often suboptimal, either depending excessively on complete attribute labeling or disregarding these attributes entirely. To overcome these limitations, we propose FairNet, a novel framework for dynamic, instance-level fairness correction. FairNet integrates a bias detector with conditional low-rank adaptation (LoRA), which enables selective activation of the fairness correction mechanism exclusively for instances identified as biased, and thereby preserve performance on unbiased instances. A key contribution is a new contrastive loss function for training the LoRA module, specifically designed to minimize intra-class representation disparities across different sensitive groups and effectively address underfitting in minority groups. The FairNet framework can flexibly handle scenarios with complete, partial, or entirely absent sensitive attribute labels. Theoretical analysis confirms that, under moderate TPR/FPR for the bias detector, FairNet can enhance the performance of the worst group without diminishing overall model performance, and potentially yield slight performance improvements. Comprehensive empirical evaluations across diverse vision and language benchmarks validate the effectiveness of FairNet.