超越人口 parity 的链接预测公平性：一项可复现性研究
Fairness in Link Prediction Beyond Demographic Parity: A Reproducibility Study
浏览论文内容
中文总结 AI 辅助
本研究复现并验证了 Mattos 等人关于人口 parity（Δ_DP）无法检测链接预测曝光偏差的观点，提出 NDKL 可检测此类偏差，复现 MORAL 的有效性并评估其鲁棒性，证实 MORAL 能减少隐藏偏差且效用损失极小。
中文摘要 AI 辅助
在公平的排名链接预测中，人口 parity（Δ_DP）是一种常用的公平性度量指标。然而，Mattos 等人（2025）指出，该指标无法检测曝光偏差，因为它忽略了链接在排名中的位置。本研究通过以下方式复现这一观点：证明当某些子组对链接被系统地排在其他链接之后时，Δ_DP 仍可指示总体 parity。不过，所提出的感知排名的归一化折扣 KL 散度（NDKL）能够检测到此类差异。我们还复现了 MORAL 的有效性，MORAL 是一种后处理方法，可在保持竞争力的效用的同时改善基于曝光的公平性。除了复现之外，我们还使用合成同质性设置、分类敏感属性以及额外的公平性和效用度量（包括适配子组对的注意力加权排名公平性（AWRF））评估了鲁棒性。总体而言，我们的结果表明，基于曝光的度量能够揭示被 Δ_DP 隐藏的偏差，而 MORAL 在不同设置和数据集下以最小的效用损失减少了这些偏差。我们在该 https URL 发布了经修正的可复现实现。
英文摘要
In fair ranked link prediction, demographic parity ($Δ_\mathrm{DP}$) is a common fairness metric. Yet, Mattos et al. (2025) argue that it fails to detect exposure bias because it ignores where links appear in the ranking. In this study, we reproduce this claim by showing that $Δ_\mathrm{DP}$ can indicate aggregate parity even when some subgroup-pair links are systematically ranked lower than others. The proposed rank-aware Normalized Discounted KL-divergence (NDKL), however, does detect such disparities. We also reproduce the effectiveness of MORAL, a post-processing method that improves exposure-based fairness while maintaining competitive utility. Beyond reproduction, we assess robustness using synthetic homophily settings, categorical sensitive attributes, and additional fairness and utility metrics, including subgroup-pair-adapted Attention-Weighted Rank Fairness (AWRF). Overall, our results show that exposure-based metrics uncover biases hidden by $Δ_\mathrm{DP}$ and that MORAL reduces these biases with minimal utility loss across diverse settings and datasets. We release a corrected, reproducible implementation at https://github.com/Floris93100/reproducing-MORAL.
发表机构
- University of Amsterdam（阿姆斯特丹大学）
机构由 AI 辅助整理，请以论文原文为准。
