注:本文来自机器之心的PaperWeekly系列:万字综述之生成对抗网络(GAN),如有侵权,请联系删除,谢谢! 前阵子学习 GAN 的过程发现现在的 GAN 综述文章大都是 2016 年 Ian Goodfellow 或者自动化所王飞跃老师那篇.可是在深度学习,GAN领域,其进展都是以月来计算的,感觉那两篇综述有些老了.最近发现有一篇最新的 GAN 综述论文(How Generative Adversarial Networks and Their Variants Work: An Over
GAN由论文<Ian Goodfellow et al., “Generative Adversarial Networks,” arXiv (2014)>提出. GAN与VAEs的区别 GANs require differentiation through the visible units, and thus cannot model discrete data, while VAEs require differentiation through the hidden units, a