arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2302.06322stat.ML
收藏

单次联邦保形预测

One-Shot Federated Conformal Prediction

Pierre Humbert, Batiste Le Bars, Aurélien Bellet, Sylvain Arlot

AI总结:

本文提出一种用于单次联邦学习的保形预测方法,通过分位数之分位数估计器在单轮通信中实现目标覆盖率,并提供本地差分隐私版本,实验证明其效果与集中式设置相近。

摘要:

在本文中,我们引入了一种保形预测方法,用于在单次联邦学习环境中构建预测集。更具体地说,我们定义了一个分位数之分位数估计器,并证明对于任何分布,仅需一轮通信即可输出具有所需覆盖率的预测集。为缓解隐私问题,我们还描述了该估计器的本地差分隐私版本。最后,在广泛的实验中,我们表明我们的方法返回的预测集在覆盖率和长度上与集中式环境中获得的预测集非常相似。总体而言,这些结果表明我们的方法特别适合在单次联邦学习环境中执行保形预测。

英文摘要:

In this paper, we introduce a conformal prediction method to construct prediction sets in a oneshot federated learning setting. More specifically, we define a quantile-of-quantiles estimator and prove that for any distribution, it is possible to output prediction sets with desired coverage in only one round of communication. To mitigate privacy issues, we also describe a locally differentially private version of our estimator. Finally, over a wide range of experiments, we show that our method returns prediction sets with coverage and length very similar to those obtained in a centralized setting. Overall, these results demonstrate that our method is particularly well-suited to perform conformal predictions in a one-shot federated learning setting.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑