Augmented Normalizing Flows: Bridging the Gap Between Generative Flows and Latent Variable Models
Machine Learning
2020-02-19 v1 Machine Learning
Abstract
In this work, we propose a new family of generative flows on an augmented data space, with an aim to improve expressivity without drastically increasing the computational cost of sampling and evaluation of a lower bound on the likelihood. Theoretically, we prove the proposed flow can approximate a Hamiltonian ODE as a universal transport map. Empirically, we demonstrate state-of-the-art performance on standard benchmarks of flow-based generative modeling.
Keywords
Cite
@article{arxiv.2002.07101,
title = {Augmented Normalizing Flows: Bridging the Gap Between Generative Flows and Latent Variable Models},
author = {Chin-Wei Huang and Laurent Dinh and Aaron Courville},
journal= {arXiv preprint arXiv:2002.07101},
year = {2020}
}
Comments
27 pages, 12 figures