Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2511.07032 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866915609475809280 |
|---|---|
| author | Zhang, Yixuan Luo, Jiabin Wang, Zhenggang Zhou, Feng Kong, Quyu |
| author_facet | Zhang, Yixuan Luo, Jiabin Wang, Zhenggang Zhou, Feng Kong, Quyu |
| contents | Fairness concerns are increasingly critical as machine learning models are deployed in high-stakes applications. While existing fairness-aware methods typically intervene at the model level, they often suffer from high computational costs, limited scalability, and poor generalization. To address these challenges, we propose a Bayesian data selection framework that ensures fairness by aligning group-specific posterior distributions of model parameters and sample weights with a shared central distribution. Our framework supports flexible alignment via various distributional discrepancy measures, including Wasserstein distance, maximum mean discrepancy, and $f$-divergence, allowing geometry-aware control without imposing explicit fairness constraints. This data-centric approach mitigates group-specific biases in training data and improves fairness in downstream tasks, with theoretical guarantees. Experiments on benchmark datasets show that our method consistently outperforms existing data selection and model-based fairness methods in both fairness and accuracy. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2511_07032 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Fair Bayesian Data Selection via Generalized Discrepancy Measures Zhang, Yixuan Luo, Jiabin Wang, Zhenggang Zhou, Feng Kong, Quyu Machine Learning Fairness concerns are increasingly critical as machine learning models are deployed in high-stakes applications. While existing fairness-aware methods typically intervene at the model level, they often suffer from high computational costs, limited scalability, and poor generalization. To address these challenges, we propose a Bayesian data selection framework that ensures fairness by aligning group-specific posterior distributions of model parameters and sample weights with a shared central distribution. Our framework supports flexible alignment via various distributional discrepancy measures, including Wasserstein distance, maximum mean discrepancy, and $f$-divergence, allowing geometry-aware control without imposing explicit fairness constraints. This data-centric approach mitigates group-specific biases in training data and improves fairness in downstream tasks, with theoretical guarantees. Experiments on benchmark datasets show that our method consistently outperforms existing data selection and model-based fairness methods in both fairness and accuracy. |
| title | Fair Bayesian Data Selection via Generalized Discrepancy Measures |
| topic | Machine Learning |
| url | https://arxiv.org/abs/2511.07032 |