跳到正文
研踪PaperTrail
← 研踪 PaperTrail

arXiv:2606.00419 · v5

Parameter-Free and Group Conditional Online Conformal Prediction

首次提交 2026-05-30 07:15
版本更新 2026-11-16 21:42
雷达观测 2026-11-17 17:11
arXiv ↗PDF ↗

摘要

Quantifying the uncertainty in predictions made by machine learning models is challenging in online settings where the characteristics of the data and labels change over time. Online conformal prediction (OCP) methods tackle this problem by producing intervals for each observation and guaranteeing that a proportion of them contain their labels, a property known as marginal coverage. OCP methods, however, may systematically miscover observations belonging to groups defined by attributes like age or race, raising fairness concerns. Group-conditional OCP (G-OCP) methods address this by producing intervals that provide coverage for multiple groups. Existing approaches rely on learning rates, which control how they react to changes in the data stream, but can be hard to tune in online settings. Leveraging advancements in learning-rate-free methods, we develop a learning-rate-free G-OCP algorithm with strong group-coverage guarantees that hold under weaker assumptions than those required in prior work. Empirical results show that our method provides group coverage, produces interval lengths competitive with those of standard methods, and reacts more effectively to changes in the data than fixed-learning-rate methods.

AI 论文解读

基于摘要 · 非全文审阅

此版本尚未生成 AI 解读。解读由后台任务生成并按论文版本保存,登录后共享已有结果。

版本差异

本站首次收录为 v5,暂无更早版本可对照。

观测到的版本

相邻记录可能跳过中间版本;首次发现 v2 及以上也会标为版本更新。

v5 · 2026-11-16 21:42 · 最新观测版本

Parameter-Free and Group Conditional Online Conformal Prediction

Quantifying the uncertainty in predictions made by machine learning models is challenging in online settings where the characteristics of the data and labels change over time. Online conformal prediction (OCP) methods tackle this problem by producing intervals for each observation and guaranteeing that a proportion of them contain their labels, a property known as marginal coverage. OCP methods, however, may systematically miscover observations belonging to groups defined by attributes like age or race, raising fairness concerns. Group-conditional OCP (G-OCP) methods address this by producing intervals that provide coverage for multiple groups. Existing approaches rely on learning rates, which control how they react to changes in the data stream, but can be hard to tune in online settings. Leveraging advancements in learning-rate-free methods, we develop a learning-rate-free G-OCP algorithm with strong group-coverage guarantees that hold under weaker assumptions than those required in prior work. Empirical results show that our method provides group coverage, produces interval lengths competitive with those of standard methods, and reacts more effectively to changes in the data than fixed-learning-rate methods.

查看此版本解读 →查看此版本原文 ↗