IFFair: Influence Function-driven Sample Reweighting for Fair Classification

2025-12-08

AI Summary

Machine learning models often perpetuate or exacerbate discriminatory decisions against disadvantaged groups due to biases in training data, undermining fairness and societal welfare. To address this, we propose a dynamic sample reweighting method grounded in influence functions—requiring no architectural modifications or feature engineering. Our approach quantifies the differential influence of individual training samples on predictions across demographic groups and incorporates gradient sensitivity to enable fine-grained, group-aware weight adjustments that jointly optimize multiple fairness constraints. This work is the first to apply influence functions for sample-level influence modeling in fair learning. Empirical evaluation across multiple real-world datasets demonstrates significant reductions in bias metrics—including equal opportunity difference and statistical parity difference—while achieving a superior Pareto trade-off between classification accuracy and fairness compared to state-of-the-art preprocessing methods.

Technology Category

• Machine Learning: Ethics, Bias, and Fairness
• Computer Vision: Bias, Fairness & Privacy
• Philosophy and Ethics of AI: Bias, Fairness & Equity

Application Category

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

Abstract

Because machine learning has significantly improved efficiency and convenience in the society, it's increasingly used to assist or replace human decision-making. However, the data-based pattern makes related algorithms learn and even exacerbate potential bias in samples, resulting in discriminatory decisions against certain unprivileged groups, depriving them of the rights to equal treatment, thus damaging the social well-being and hindering the development of related applications. Therefore, we propose a pre-processing method IFFair based on the influence function. Compared with other fairness optimization approaches, IFFair only uses the influence disparity of training samples on different groups as a guidance to dynamically adjust the sample weights during training without modifying the network structure, data features and decision boundaries. To evaluate the validity of IFFair, we conduct experiments on multiple real-world datasets and metrics. The experimental results show that our approach mitigates bias of multiple accepted metrics in the classification setting, including demographic parity, equalized odds, equality of opportunity and error rate parity without conflicts. It also demonstrates that IFFair achieves better trade-off between multiple utility and fairness metrics compared with previous pre-processing methods.

Problem

• Mitigates algorithmic bias in classification tasks
• Uses influence functions to reweight training samples
• Balances fairness and utility without structural changes

Innovation

• Influence function-driven sample reweighting for fairness
• Dynamic weight adjustment without altering model structure
• Pre-processing method balancing utility and fairness metrics