SkillFeeder › 全部 Skill › 数据与复盘 › auditing-subgroup-fairness

审计医疗AI模型在性别、年龄、种族等亚组上的识别与脱敏公平性

auditing-subgroup-fairness

auditing-subgroup-fairness 是 SkillFeeder 收录的Agent Skill,来自 GitHub 仓库 maziyarpanahi/openmed,★ 5,783。审计医疗AI模型在性别、年龄、种族等亚组上的识别与脱敏公平性。适合临床AI治理者、医疗NLP工程师。场景是数据与复盘。能做的事包括输出各亚组的召回率与PHI泄露率、自动标记未覆盖亚组的数据缺口、生成可交付的公平性评估报告。打开 GitHub 自行安装。

打开 GitHub

仓库
maziyarpanahi/openmed
仓库星标
★ 5,783
场景
数据与复盘
适合
临床AI治理者、医疗NLP工程师
类型
Skill

能做什么

原文简介

Audit an OpenMed NER or de-identification model for performance disparities across demographic subgroups (sex, age band, race/ethnicity when available) using openmed.eval.fairness_report. Use when the user wants per-subgroup recall and leakage, wants to check whether de-identification under-protects a group, wants to surface a documentation gap where subgroup data is missing, or needs equalized-od

同场景推荐