---
title: "高召回候选生成阶段的有限样本覆盖审计：认证与学习理论设计"
source: "arXiv AI/CL/LG（论文）"
source_tier: "t2"
url: "https://arxiv.org/abs/2607.21480v1"
permalink: "https://feed.astralor.com/items/itm_3b82917c630df5b48e5da48e"
published: "2026-07-23T16:21:30.000Z"
score: 78
channel: "ai"
tags: ["候选生成","审计","有限样本","认证"]
---

# 高召回候选生成阶段的有限样本覆盖审计：认证与学习理论设计

## AI 点评

> 这篇论文的核心贡献在于从理论层面揭示了高召回候选集审计的标复杂度本质，特别是证明了“仅靠点中查看候选集内部无法认证遗漏”这一反直觉结论，对实际系统设计有重要指导意义。对于从事法律电子发现、医疗影像筛查等严格召回率要求的从业者来说，本文提供的精确有限样本工具包将极大提升审计的可信度和效率。我建议所有关注AI系统可靠性评估的研究者和工程师仔细阅读。

## 速览

本文研究了高召回候选生成阶段的审计标签复杂度问题，证明了仅从候选集内采样无法认证遗漏的相关项数量，必须从排除池中抽样。作者推导了零遗漏情况下的极小极大最优下界，并开发了一套基于二项和超几何逆变换的精确有限样本工具包，可认证遗漏质量、转换为召回率，并同时认证嵌套候选生成器族，还能对声明的扰动机制进行压力测试。所有保证成立的条件是候选生成器（或其预指定族）和审计规则在检查认证标签前固定。

## 关键点

- 任何仅使用候选集内标签的审计程序都无法有效认证遗漏的相关项数量，必须从排除池中抽样。
- 零遗漏情况下，有效审计所需排除池标签数量的下界为 N0/m 量级，且该策略是极小极大最优的。
- 开发了基于二项和超几何逆变换的精确有限样本工具包，支持遗漏质量认证、召回率转换、嵌套生成器同时认证及压力测试。
- 所有保证依赖于候选生成器（或预指定族）和审计规则在查看标签前固定，避免数据窥探偏差。

## 全文

An initial high-recall stage in an empirical pipeline decides which items pass to later review, labelling, or modelling, and relevant items it misses are lost to every subsequent stage. We study how many audit labels are needed to certify, with finite-sample validity, that this missed relevant mass is small, and our main results characterise the label complexity of this problem. We first show that no procedure using only labels from inside the candidate set can certify any non-trivial bound on the missed mass: the audit must sample the excluded pool, the only region where unrecovered relevant items can lie. We then prove a matching finite-corpus lower bound. Any valid audit that certifies fewer than $m$ missed relevant items with high probability when none are present, even if adaptive and permitted to label the entire included pool, must inspect on the order of $N_0/m$ excluded-pool labels. Excluded-pool auditing is therefore minimax rate-optimal, not merely convenient, for missed-mass certification in the zero-miss regime. Building on this characterisation, we develop an exact finite-sample toolkit, using binomial and hypergeometric inversion rather than asymptotic approximation, that certifies missed mass, converts it to recall through a two-pool design, certifies pre-specified families of nested candidate generators simultaneously, and produces stress-test certificates against declared perturbation mechanisms. These certificates can be paired with observable review burden to select the least burdensome pre-specified candidate generator meeting a missed-mass target. Every guarantee holds under one discipline: the candidate generator, or the pre-specified family from which it is selected, and the audit rule are fixed before the certification labels are examined.

---

引用前请核对原文;本文由星际流动整理:https://feed.astralor.com/items/itm_3b82917c630df5b48e5da48e