Header logo is

Adversarial Variational Bayes: Unifying Variational Autoencoders and Generative Adversarial Networks

2017

Conference Paper

avg


Variational Autoencoders (VAEs) are expressive latent variable models that can be used to learn complex probability distributions from training data. However, the quality of the resulting model crucially relies on the expressiveness of the inference model. We introduce Adversarial Variational Bayes (AVB), a technique for training Variational Autoencoders with arbitrarily expressive inference models. We achieve this by introducing an auxiliary discriminative network that allows to rephrase the maximum-likelihood-problem as a two-player game, hence establishing a principled connection between VAEs and Generative Adversarial Networks (GANs). We show that in the nonparametric limit our method yields an exact maximum-likelihood assignment for the parameters of the generative model, as well as the exact posterior distribution over the latent variables given an observation. Contrary to competing approaches which combine VAEs with GANs, our approach has a clear theoretical justification, retains most advantages of standard Variational Autoencoders and is easy to implement.

Author(s): L. Mescheder and S. Nowozin and A. Geiger
Book Title: International Conference on Machine Learning (ICML) 2017
Year: 2017
Month: August

Department(s): Autonomous Vision
Bibtex Type: Conference Paper (inproceedings)
Paper Type: Conference

Event Name: International Conference on Machine Learning (ICML) 2017
Event Place: Sydney, Australia

Links: pdf
suppmat
Project Page
arxiv-version

BibTex

@inproceedings{Mescheder2017ICML,
  title = {Adversarial Variational Bayes: Unifying Variational Autoencoders and Generative Adversarial Networks},
  author = {Mescheder, L. and Nowozin, S. and Geiger, A.},
  booktitle = {International Conference on Machine Learning (ICML) 2017},
  month = aug,
  year = {2017},
  month_numeric = {8}
}