BriefGPT.xyz
Jan, 2017
训练生成对抗网络的原则性方法探究
Towards Principled Methods for Training Generative Adversarial Networks
HTML
PDF
Martin Arjovsky, Léon Bottou
TL;DR
本文旨在全面理解对抗生成网络训练动态，研究了训练中出现的不稳定性和饱和现象，提出了一种实用且基于理论的解决方向，并引入了新的工具来研究这些问题。
Abstract
The goal of this paper is not to introduce a single algorithm or method, but to make theoretical steps towards fully understanding the
training dynamics
of
generative adversarial networks
. In order to substantiat
→