Wasserstein Auto-Encoders: Latent Dimensionality and Random Encoders

Paul K. Rubenstein, Bernhard Schoelkopf, Ilya Tolstikhin

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.

Keywords: wasserstein auto-encoders, WAE

TL;DR: We study the role of latent dimensionality in Wasserstein Auto-Encoders, and show that random encoders may often be preferable to deterministic encoders.