Comparison of Maximum Likelihood and GAN-based training of Real NVPs

Danihelka, I; Lakshminarayanan, B; Uria, B; Wierstra, D; Dayan, P

doi:10.48550/arXiv.1705.05263

Item

ITEM ACTIONSEXPORT

Add to Basket

Local TagsRelease HistoryDetailsSummary

Released

Preprint

Comparison of Maximum Likelihood and GAN-based training of Real NVPs

MPS-Authors

There are no MPG-Authors in the publication available

External Resource

https://arxiv.org/abs/1705.05263
(Any fulltext)

Fulltext (restricted access)

There are currently no full texts shared for your IP range.

Fulltext (public)

There are no public fulltexts stored in PuRe

Supplementary Material (public)

There is no public supplementary material available

Citation

Danihelka, I., Lakshminarayanan, B., Uria, B., Wierstra, D., & Dayan, P. (submitted). Comparison of Maximum Likelihood and GAN-based training of Real NVPs.

Cite as: https://hdl.handle.net/21.11116/0000-0002-6857-8

Abstract

We train a generator by maximum likelihood and we also train the same generator architecture by Wasserstein GAN. We then compare the generated samples, exact log-probability densities and approximate Wasserstein distances. We show that an independent critic trained to approximate Wasserstein distance between the validation set and the generator distribution helps detect overfitting. Finally, we use ideas from the one-shot learning literature to develop a novel fast learning critic.