Spatial Variational Auto-Encoding via Matrix-Variate Normal Distributions

Abstract

The key idea of variational auto-encoders (VAEs) resembles that of traditional auto-encoder models in which spatial information is supposed to be explicitly encoded in the latent space. However, the latent variables in VAEs are vectors, which are commonly interpreted as multiple feature maps of size 1x1. Such representations can only convey spatial… (More)

Topics

4 Figures and Tables

Slides referencing similar topics