---
title: "KAISEN：面向临床风险模型的子组公平性审计流水线"
source: "arXiv AI/CL/LG（论文）"
source_tier: "t2"
url: "https://arxiv.org/abs/2607.28608v1"
permalink: "https://feed.astralor.com/items/itm_6059c4ea9e78541a23705ac1"
published: "2026-07-30T17:57:18.000Z"
score: 73
channel: "ai"
tags: ["CUSUM","Equalized Odds","KAISEN","临床风险模型","公平性审计"]
---

# KAISEN：面向临床风险模型的子组公平性审计流水线

## AI 点评

> 这篇论文最值得看的地方在于它把审计工具本身“放到砧板上”做压力测试——结果那些看似合理的审计组件可能在关键时刻静默失败。对任何负责任地做临床AI评估的团队来说，这都应该是必读清单的一部分。另外，它强调“报告方差而非平均值”这个观点，简直是对当前大量公平性报告的一记当头棒喝。

## 速览

KAISEN是面向临床风险模型的可复现子组公平性审计流水线，包含子组分层、差异测量、机制诊断、事后缓解和漂移监测五个阶段。基于16种疾病任务、15个健康社会决定因素轴和3个预定交叉组的合成基准，研究发现：显著性计数与原始均等优势差异（EOD）的相关性需经最小可检测效应标准化后才显著；按组阈值优化在48次中全部降低EOD，而Platt缩放表现类似抛硬币；机制诊断在代理设定错误时可能完全失效但不露痕迹；CUSUM漂移监测的阈值跨队列迁移失败。结果表明审计需报告方差而非平均值。

## 关键点

- KAISEN包含五阶段审计流水线，并在合成基准上测试至失败点；
- 按组阈值优化在48次保留运行中100%降低均等优势差异，而Platt缩放仅改善19/48次；
- 机制诊断在代理变量设定错误时有144/144受控案例正确，但无法恢复任何模型驱动案例且无失败信号；
- CUSUM漂移检测的参考阈值在跨队列时产生27次全部误报和7/8次漏报，显示阈值不可迁移。

## 全文

Clinical risk models routinely achieve strong aggregate performance while producing materially different error rates across patient subgroups. Audit pipelines have been proposed to catch this, but their components are rarely stress-tested, so it is unclear which parts of an audit can be trusted and under what conditions. We present KAISEN, a five-phase audit pipeline covering subgroup stratification, disparity measurement, mechanism diagnostics, post-hoc mitigation, and drift monitoring, evaluated to the point of failure on a synthetic benchmark of 16 disease tasks, 15 social-determinant axes from Healthy People 2030, and three prespecified intersections. Four findings follow. (i) Significance tracks each axis's gap against its own minimum detectable effect: rank correlation between significance count and raw equalized-odds difference (EOD) across the 15 axes is rho = 0.56, rising to rho = 0.78 once EOD is standardized by that floor. (ii) Per-group threshold optimization reduces EOD in 48 of 48 held-out runs (paired delta = -0.285, 95% CI [-0.313, -0.252]), while group-wise Platt scaling -- the better calibrator -- behaves as a coin flip on EOD (19 of 48 runs improved, 95% CI [0.26, 0.55]) with mean effect near zero, so what an audit should report is the variance, not the average. (iii) The mechanism diagnostic classifies 144 of 144 controlled cases correctly but recovers none of 48 model-driven cases under proxy misspecification, with no signal that it failed. (iv) CUSUM failures and false alarms track cohort realization far more than disease: at the reference threshold, all 27 false alarms and 7 of 8 missed shifts come from different seeds (chi-squared p = 0.002), so a threshold tuned on one cohort fails to transfer. All results are synthetic with known ground truth and do not establish clinical validity. Code, artifacts, and scripts reproducing every number are released.

---

引用前请核对原文;本文由星际流动整理:https://feed.astralor.com/items/itm_6059c4ea9e78541a23705ac1