BriefGPT.xyz
Mar, 2018
关于生成对抗网络的一些理论性质
Some Theoretical Properties of GANs
HTML
PDF
G. Biau, B. Cadre, M. Sangnier, U. Tanielian
TL;DR
本文通过分析数学和统计学特性，研究生成式对抗网络与Jensen-Shannon散度的深刻联系，并提供鉴别器族群的逼近论证及样本估计分布的大样本性质，特别地证明了中心极限定理。
Abstract
generative adversarial networks
(
gans
) are a class of generative algorithms that have been shown to produce state-of-the art samples, especially in the domain of
→