English(EN)When Fairness Metrics Disagree: Evaluating the Reliability of Demographic Fairness Assessment in Machine Learning
新研究聚焦扩散模型、朴素贝叶斯和空间模式中的人工智能公平性
作者PulseAugur 编辑部·[11 个来源]·
研究人员正在开发新方法，以确保各种应用中机器学习模型的公平性。一篇论文介绍了“StayFair”，通过将偏差分解为模型和引导分量，来在不同引导尺度下保持扩散模型的公平性。另一项研究提出了一种“偏差缓解朴素贝叶斯”分类器，该分类器融合了特定群体和汇总的似然估计，以平衡公平性和准确性。此外，一种新方法基于个体移动模式评估空间公平性，将该概念推广到静态位置之外。其他研究探讨了不同公平性指标之间的不一致性，强调了多指标分析的必要性，并利用最优传输方法提出了在感知和非感知设置下公平回归的统一框架。
AI
arXiv:2605.28036v1 Announce Type: cross Abstract: Diffusion models steer conditional generation with a tunable guidance scale to trade off prompt alignment and diversity. However, existing debiasing techniques are optimized for a single scale, degrading fairness when users adjust…
arXiv cs.LG
TIER_1English(EN)·John Arthur Junior, Abdul Lateef Yussif, Maame G. Asante-Mensah, Charles R. Haruna, Sandro Amofa, Elliot Attipoe·
arXiv:2605.25228v1 Announce Type: new Abstract: Concerns about algorithmic bias and fairness have increased as artificial intelligence has been incorporated into high-stakes decision-making. Traditional Naive Bayes classifiers, while efficient and interpretable, lack fairness-awa…
arXiv cs.LG
TIER_1English(EN)·Francesco Lettich, Mario A. Nascimento, Chiara Pugliese, Chiara Renso·
arXiv:2605.23234v1 Announce Type: new Abstract: Assessing the spatial fairness of predictive models involves establishing whether they are statistically penalizing (favoring) individuals associated with certain geographical locations. Literature on this topic makes the fundamenta…
arXiv:2604.15038v2 Announce Type: replace-cross Abstract: The evaluation of fairness in machine learning systems has become a central concern in high-stakes applications, including biometric recognition, healthcare decision-making, and automated risk assessment. Existing approach…
arXiv stat.ML
TIER_1English(EN)·M. Generali Lince, V. Divol, R. Flamary, S. Gaucher, P. Loiseau·
arXiv:2605.28233v1 Announce Type: new Abstract: Fairness-accuracy trade-offs are a central concern in the deployment of fairness-aware machine learning methods. When sensitive attributes are unavailable at inference time-the so called unawareness setting, principled methods for o…
arXiv stat.ML
TIER_1English(EN)·M. Generali Lince, S. Gaucher, J-J. Vie, P. Loiseau·
arXiv:2605.28251v1 Announce Type: new Abstract: We consider the problem of learning a counterfactually fair regressor. We adopt a causal uncertainty view in which counterfactual fairness is defined with resampled noise. We focus on obtaining theoretical fairness guarantees for a …
We consider the problem of learning a counterfactually fair regressor. We adopt a causal uncertainty view in which counterfactual fairness is defined with resampled noise. We focus on obtaining theoretical fairness guarantees for a new post-processing estimator. We begin by showi…
Fairness-accuracy trade-offs are a central concern in the deployment of fairness-aware machine learning methods. When sensitive attributes are unavailable at inference time-the so called unawareness setting, principled methods for obtaining accurate predictions under relaxed fair…
Diffusion models steer conditional generation with a tunable guidance scale to trade off prompt alignment and diversity. However, existing debiasing techniques are optimized for a single scale, degrading fairness when users adjust this parameter. We trace this behavior to a previ…
arXiv:2605.23145v1 Announce Type: new Abstract: Individual fairness, the notion that "similar individuals should be treated similarly," provides a strong and flexible fairness guarantee for algorithmic decision makers. However, a barrier to implementing individual fairness in pra…
Individual fairness, the notion that "similar individuals should be treated similarly," provides a strong and flexible fairness guarantee for algorithmic decision makers. However, a barrier to implementing individual fairness in practice is the difficulty of learning the similari…
