#Implementation of Adversarial Autoenocder (arXiv : http://arxiv.org/abs/1511.05644)

Python codes for training Adversarial Autoenocder with Theano.

You can train example models of Adversarial Autoencoder for MNIST on the ipython notebook.

See https://github.com/takerum/adversarial_autoencoder/blob/master/train_adv_autoenc.ipynb .

• The hidden code z of the trained adversarial autoencoder with 2-D uniform prior distribution.

• The input x sampled from the trained adversarial autoencoder with 8-D gaussian prior distribution.

##Required libraries: python 2.7, theano 0.7.0

##References: [1] Alireza Makhzani, Jonathon Shlens, Navdeep Jaitly and Ian Goodfellow. Adversarial Autoenocders. arXiv preprint (http://arxiv.org/abs/1511.05644).