On the Latent Space of Wasserstein Auto-Encoders

Abstract

We study the role of latent space dimensionality in Wasserstein auto-encoders (WAEs). Through experimentation on synthetic and real datasets, we argue that random encoders should be preferred over deterministic encoders. We highlight the potential of WAEs for representation learning with promising results on a benchmark disentanglement task.

References in corpus (2)

• Auto-Encoding Variational Bayes
• GANs Trained by a Two Time-Scale Update Rule Converge to a Local Nash Equilibrium

Cited by in corpus (12)

• Learning Factorized Multimodal Representations
• Educating Text Autoencoders: Latent Representation Guidance via Denoising
• Poincaré Wasserstein Autoencoder
• Disentangled Recurrent Wasserstein Autoencoder
• Perceptual Generative Autoencoders
• Autoencoding topology
• Training Generative Reversible Networks
• A New Framework for Query Efficient Active Imitation Learning
• Variation Network: Learning High-level Attributes for Controlled Input Manipulation
• Cauchy-Schwarz Regularized Autoencoder
• Topic Modeling with Wasserstein Autoencoders
• Learning Deep-Latent Hierarchies by Stacking Wasserstein Autoencoders