BriefGPT.xyz
Sep, 2023
用多项式数量的样本来隐私学习高斯混合模型
Mixtures of Gaussians are Privately Learnable with a Polynomial Number of Samples
HTML
PDF
Mohammad Afzali, Hassan Ashtiani, Christopher Liaw
TL;DR
在满足差分隐私的约束下，研究了估计混合高斯模型问题。通过使用新的框架，证明了高斯模型类的混合模型是可私密学习的，得到了估计混合高斯模型所需的样本数量的有界性，且不对GMMs作出任何结构性假设。
Abstract
We study the problem of
estimating mixtures
of
gaussians
under the constraint of
differential privacy
(DP). Our main result is that $\tild
→