On the Latent Space of Wasserstein Auto-Encoders

2018/02/11 by Paul K. Rubenstein, Rubenstein, Paul K., Bernhard Schoelkopf +3 · 1 citation
Computer Science · Mathematics · #FOS: Computer and information sciences #Machine Learning (cs.LG) #Machine Learning (stat.ML) #cs.LG #stat.ML

paper · pdf · doi:10.48550/arxiv.1802.03761

arxiv created 2018/02/11 · arxiv updated 2018/02/13

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.

Cited by

• History Matching under Uncertainty of Geological Scenarios with Implicit Geological Realism Control with Generative Deep Learning and Graph Convolutions

Related

• Wasserstein Auto-Encoders
• Statistical Regeneration Guarantees of the Wasserstein Autoencoder with Latent Space Consistency
• A Statistical Analysis of Wasserstein Autoencoders for Intrinsically Low-dimensional Data
• Wasserstein-Wasserstein Auto-Encoders
• Symmetric Wasserstein Autoencoders