Adversarial Autoencoders (AAE)

This is a Tensorflow 2.0 implementation of Adversarial Autoencoders by Alireza Makhzani et al. (ICLR 2016). This repository contains reproduce of several experiments mentioned in the paper.

Requirements

• Python 3
• TensorFlow 2.0+
• Numpy
• Matplotlib

Results

Unsupervised AAE deterministic

Latent space Reconstruction Sampled

Unsupervised AAE deterministic convolutional

Latent space Reconstruction Sampled

Unsupervised AAE deterministic convolutional using WGAN loss function

Latent space Reconstruction Sampled

Unsupervised AAE probabilistic

Latent space Reconstruction Sampled

Unsupervised AAE probabilistic convolutional

Latent space Reconstruction Sampled

Supervised AAE deterministic

Latent space Reconstruction Sampled

Supervised AAE deterministic convolutional

Latent space Reconstruction Sampled