Adversarial Autoencoders
This work addresses the challenge of effective variational inference and generative modeling for researchers and practitioners in machine learning, though it is incremental as it builds on existing GAN and autoencoder frameworks.
The paper tackles the problem of learning deep generative models by proposing adversarial autoencoders (AAE), which use generative adversarial networks to match the aggregated posterior of an autoencoder's hidden code to a prior distribution, resulting in competitive performance on tasks like generative modeling and semi-supervised classification across datasets such as MNIST, Street View House Numbers, and Toronto Face.
In this paper, we propose the "adversarial autoencoder" (AAE), which is a probabilistic autoencoder that uses the recently proposed generative adversarial networks (GAN) to perform variational inference by matching the aggregated posterior of the hidden code vector of the autoencoder with an arbitrary prior distribution. Matching the aggregated posterior to the prior ensures that generating from any part of prior space results in meaningful samples. As a result, the decoder of the adversarial autoencoder learns a deep generative model that maps the imposed prior to the data distribution. We show how the adversarial autoencoder can be used in applications such as semi-supervised classification, disentangling style and content of images, unsupervised clustering, dimensionality reduction and data visualization. We performed experiments on MNIST, Street View House Numbers and Toronto Face datasets and show that adversarial autoencoders achieve competitive results in generative modeling and semi-supervised classification tasks.