语义特征改变、操作:
原视频 https://pan.baidu.com/s/1kV0kbS7
2.3
An appealing property of GAN is that its discriminator net- work implicitly has to learn a rich similarity metric for im- ages, so as to discriminate them from “non-images”. We thus propose to exploit this observation so as to transfer the properties of images learned by the discriminator into a more abstract reconstruction error for the VAE. The end re- sult will be a method that combines the advantage of GAN as a high quality generative model and VAE as a method that produces an encoder of data into the latent space z.
GAN的判别网络需要学习图片中的丰富特征。
因此整合GAN的高质量生成模型优点和VAE的隐变量学习
Specifically, since element-wise reconstruction errors are not adequate for images and other signals with invariances, we propose replacing the VAE reconstruction (expected log likelihood) error term from Eq. 3 with a reconstruction er- ror expressed in the GAN discriminator. To achieve this, let Disl(x) denote the hidden representation of the lth layer of the discriminator. We introduce a Gaussian observation model for Disl(x) with mean Disl(x ̃) and identity covari- ance:
VAE的生成使用GAN。 整合了VAE和GAN。
论文主要内容:
论文贡献:
https://arxiv.org/abs/1512.09300
论文代码:https://github.com/andersbll/autoencoding_beyond_pixels
其他相关代码:https://github.com/timsainb/Tensorflow-MultiGPU-VAE-GAN
https://github.com/staturecrane/dcgan_vae_torch
本文由zdx3578推荐。