在最早的对抗式生成网络的论文中,生成器和分类器用的都是全联接层,所以没有办法很好的生成图片数据,也没有办法做的很深。所以在随后的论文中,人们提出了深度卷积对抗式生成网络(deep convolutional generative adversarial network or DCGAN)\[[2](#参考文献)\]。在DCGAN中,生成器 G 是由多个卷积转置层(transposed convolution)组成的,这样可以用更少的参数来生成质量更高的图片。具体网络结果可参见图3。
由于上面的这种网络都是由全联接层组成,所以没有办法很好的生成图片数据,也没有办法做的很深。所以在随后的论文中,人们提出了深度卷积对抗式生成网络(deep convolutional generative adversarial network or DCGAN)\[[2](#参考文献)\]。在DCGAN中,生成器 G 是由多个卷积转置层(transposed convolution)组成的,这样可以用更少的参数来生成质量更高的图片。具体网络结果可参见图4。而判别器是由多个卷积层组成。
1. Goodfellow I, Pouget-Abadie J, Mirza M, et al. [Generative adversarial nets](https://arxiv.org/pdf/1406.2661v1.pdf)[C] Advances in Neural Information Processing Systems. 2014
1. Goodfellow I, Pouget-Abadie J, Mirza M, et al. [Generative adversarial nets](https://arxiv.org/pdf/1406.2661v1.pdf)[C] Advances in Neural Information Processing Systems. 2014
2. Radford A, Metz L, Chintala S. [Unsupervised Representation Learning with Deep Convolutional Generative Adversarial Networks](https://arxiv.org/pdf/1511.06434v2.pdf)[C] arXiv preprint arXiv:1511.06434. 2015
2. Radford A, Metz L, Chintala S. [Unsupervised Representation Learning with Deep Convolutional Generative Adversarial Networks](https://arxiv.org/pdf/1511.06434v2.pdf)[C] arXiv preprint arXiv:1511.06434. 2015
3. Kingma D.P. and Welling M. [Auto-encoding variational bayes](https://arxiv.org/pdf/1312.6114v10.pdf)[C] arXiv preprint arXiv:1312.6114. 2013
3. Kingma D.P. and Welling M. [Auto-encoding variational bayes](https://arxiv.org/pdf/1312.6114v10.pdf)[C] arXiv preprint arXiv:1312.6114. 2013
4. Hinton G and Salakhutdinov R. [Reducing the dimensionality of data with neural networks](https://www.cs.toronto.edu/~hinton/science.pdf) Science 313.5786. 2006
5. Salakhutdinov R and Hinton G. [Deep Boltzmann Machines](http://www.jmlr.org/proceedings/papers/v5/salakhutdinov09a/salakhutdinov09a.pdf)[J] AISTATS. Vol. 1. 2009
6. Larochelle H and Murray I. [The Neural Autoregressive Distribution Estimator](http://www.jmlr.org/proceedings/papers/v15/larochelle11a/larochelle11a.pdf) AISTATS. Vol. 1. 2011.
7. van den Oord A, Kalchbrenner N and Kavukcuoglu K. [Pixel Recurrent Neural Networks](https://arxiv.org/pdf/1601.06759v3.pdf) arXiv preprint arXiv:1601.06759 (2016).