从分布稳定性视角统一理解公平性（arXiv 论文）
A Comprehensive View of Fairness through Distributional Stability
AI 导读
一篇 arXiv 论文提出将算法公平性视为分布稳定性：预测器在改变受保护群体构成的扰动下保持稳定即为公平，经典公平性定义可归结为对不同扰动的稳定性，不公平度差距对应预测率泛函的 Lipschitz 常数。该方法无需已知部署分布即可在测试时对一系列人口构成提供一致性保证，并给出基于重加权预测器凸组合的二阶锥规划学习程序及泛化界，在标准基准上验证了效果。
正文 · 原文
Abstract:We view fairness as a property of distributional stability. Rather than assessing a predictor under a fixed data distribution, we study how its predictions change under perturbations that modify the composition of protected groups. A predictor is fair if it remains stable under such shifts. Under this perspective, several classical notions of fairness arise as stability with respect to specific perturbations, with the associated unfairness gap given by a Lipschitz constant of a prediction-rate functional. This formulation also yields guarantees that hold uniformly over a range of demographic compositions at test time, without requiring knowledge of the deployment distribution. It leads to a learning procedure based on convex combinations of reweighted predictors, formulated as a second-order cone program, for which we establish generalization bounds. Experiments on standard benchmarks illustrate the approach.
| Subjects: | Machine Learning (cs.LG); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2609.37061 [cs.LG] |
| (or arXiv:2609.37061v2 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2609.37061 arXiv-issued DOI via DataCite |
Submission history
From: Charlotte Laclau [view email]
[v1]
Tue, 29 Sep 2026 09:00:49 UTC (361 KB)
[v2]
Fri, 30 Oct 2026 10:42:03 UTC (545 KB)
来源：arXiv cs.AI · arxiv.org