BriefGPT.xyz
Jun, 2023
带有许多类别的类条件拟合预测
Class-Conditional Conformal Prediction With Many Classes
HTML
PDF
Tiffany Ding, Anastasios N. Angelopoulos, Stephen Bates, Michael I. Jordan, Ryan J. Tibshirani
TL;DR
本文提出了一种称为聚类置信预测的方法,该方法通过聚类类,将具有“相似”信心得分的类聚合在一起,然后在聚类水平上执行置信预测,进行分类问题中的更强的置信度验证, 应用于多分类图像数据集中表现更加优秀。
Abstract
Standard
conformal prediction
methods provide a marginal coverage guarantee, which means that for a random test point, the
conformal prediction
set contains the true label with a user-chosen probability. In many
→