Elicited Adaptation: Auditable Localized Fairness via Pairwise Queries
Abstract
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 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 , and a full cost curve describes finite target-only repair budgets. Committing before the target is known can cost 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 bound, similarity queries alone yield a feasible -approximate repair for tolerance , reducing movement relative to common repair by a factor linear in . Feasible -optimal repair on this family has worst-case tolerance-query complexity 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.
Figures & tables
| Respondent panel | Grid median | Logistic median | Grid IQR | Acceptance | |
|---|---|---|---|---|---|
| Black/African-American | .035 | .039 | .40 | 42% | |
| White/Caucasian | .060 | .063 | .43 | 50% |
| Procedure | Accuracy | Strict viol. | Brier | Log loss |
|---|---|---|---|---|
| Lipschitz-U | ||||
| Lipschitz-H | ||||
| Lipschitz-M | ||||
| Threshold-U | ||||
| Threshold-H | ||||
| Threshold-M |
Algorithmic Fairness
Algorithmic fairness studies how automated decisions distribute errors, opportunities, and harms across groups, and how to measure and reduce disparities.
See all 452 papers