arXiv 论文提出可审计的局部公平性配对查询框架
Elicited Adaptation: Auditable Localized Fairness via Pairwise Queries
这篇论文把公平性标准变化后的模型修复成本做成了可计算的问题,给出了 n-1 倍承诺代价这样的硬界。做算法公平性或合规审计的研友可以看看,查询复杂度部分对设计实际征询流程也有参考价值。
一篇 arXiv 论文提出"Elicited Adaptation"框架,研究当社区给出的公平性标准(由相似性与容忍度判断定义受保护比较和可接受分数差距)发生变化时,预测器分数在 n 个画像的固定池上所需的最小均匀移动量。论文刻画了在适配过程中维持源标准的精确成本:沿最短路径的连续目标约束段决定最优放大因子,其最坏情形为 ⌊n/2⌋;若在目标标准揭晓前提前承诺,成本可达揭晓后修复最坏成本的 n-1 倍,且该界在二维 Mahalanobis 度量的凸球内依然成立。
在机制层面,论文给出正的边预算下限同时约束两个因子,有界几何结构也限制承诺成本。在达到 n-1 界的两个内部目标上,仅凭相似性查询即可在容忍度 ε≤1/[2(n-1)] 时得到可行的 2-近似修复,其移动量相对常见修复方法按 n 的线性因子减少。在该构造族上,达到 (1+α)-最优修复的容忍度查询复杂度最坏为 Θ(α⁻²n²log(1/δ))(固定斜率 logistic 模型下)。
论文配套了引导式实验与训练对比,考察容忍度选择及数值合规与预测效用的关系,将"获得显著适配收益所需的信息"与"接近最优移动所需的精度"区分开来。
Elicited Adaptation: Auditable Localized Fairness via Pairwise Queries
How much must a predictor's scores change when a community's elicited fairness standard changes? Similarity and tolerance judgments define protected comparisons and acceptable score gaps. We study minimum uniform score movement on a fixed pool of $n$ profiles. For scores satisfying a source standard, we characterize the exact cost of retaining that standard during adaptation: runs of consecutive target constraints along shortest paths determine the optimal amplification factor. Its sharp worst case is $\lfloor n/2\rfloor$, and a full cost curve describes finite target-only repair budgets. Committing before the target is known can cost $n-1$ times the worst repair after revelation, even within one fixed convex ball of two-dimensional Mahalanobis metrics. Positive edge-budget floors cap both factors; bounded geometry also limits the commitment cost. At two interior targets of the construction attaining the $n-1$ bound, similarity queries alone yield a feasible $2$-approximate repair for tolerance $\eps\le1/[2(n-1)]$, reducing movement relative to common repair by a factor linear in $n$. Feasible $(1+α)$-optimal repair on this family has worst-case tolerance-query complexity $Θ(α^{-2}n^2\log(1/δ))$ under a fixed-slope logistic model. The analysis separates the information needed for a substantial adaptation benefit from the precision needed for nearly optimal movement. Elicitation studies and training comparisons examine tolerance choice and the relation between numerical compliance and predictive utility.
This report is categorized under AITOP's "Papers" desk. Key entities and technical topics under continuous tracking include arXiv, machine learning theory.
Source attribution: curated from arXiv cs.LG (@Shrey Shah).