BriefGPT.xyz
Feb, 2017
从生成对抗网络中准确恢复潜在向量
Precise Recovery of Latent Vectors from Generative Adversarial Networks
HTML
PDF
Zachary C. Lipton, Subarna Tripathi
TL;DR
通过引入随机剪裁的梯度下降技术，我们实现了基于生成式对抗网络（GANs）生成的图像的编码的完美还原，并且该方法对噪声表现出鲁棒性，甚至可以用于恢复未知图像的编码。
Abstract
generative adversarial networks
(
gans
) transform
latent vectors
into visually plausible images. It is generally thought that the original
→