AI Hub
All papers
Architecture

Auto-Encoding Variational Bayes

University of Amsterdam·December 20, 2013

Diederik P. Kingma, Max Welling

View on arXiv

TL;DR

Introduces the variational autoencoder, a deep generative model with a smooth, samplable latent space.

Why it matters

A foundational generative model and a building block of later systems like latent diffusion.

Related terms