Geometry Score: A Method For Comparing Generative Adversarial Networks

Valentin Khrulkov, Ivan Oseledets
Proceedings of the 35th International Conference on Machine Learning, PMLR 80:2621-2629, 2018.

Abstract

One of the biggest challenges in the research of generative adversarial networks (GANs) is assessing the quality of generated samples and detecting various levels of mode collapse. In this work, we construct a novel measure of performance of a GAN by comparing geometrical properties of the underlying data manifold and the generated one, which provides both qualitative and quantitative means for evaluation. Our algorithm can be applied to datasets of an arbitrary nature and is not limited to visual data. We test the obtained metric on various real-life models and datasets and demonstrate that our method provides new insights into properties of GANs.

Cite this Paper

BibTeX
@InProceedings{pmlr-v80-khrulkov18a, title = {Geometry Score: A Method For Comparing Generative Adversarial Networks}, author = {Khrulkov, Valentin and Oseledets, Ivan}, booktitle = {Proceedings of the 35th International Conference on Machine Learning}, pages = {2621--2629}, year = {2018}, editor = {Dy, Jennifer and Krause, Andreas}, volume = {80}, series = {Proceedings of Machine Learning Research}, month = {10--15 Jul}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v80/khrulkov18a/khrulkov18a.pdf}, url = {https://proceedings.mlr.press/v80/khrulkov18a.html}, abstract = {One of the biggest challenges in the research of generative adversarial networks (GANs) is assessing the quality of generated samples and detecting various levels of mode collapse. In this work, we construct a novel measure of performance of a GAN by comparing geometrical properties of the underlying data manifold and the generated one, which provides both qualitative and quantitative means for evaluation. Our algorithm can be applied to datasets of an arbitrary nature and is not limited to visual data. We test the obtained metric on various real-life models and datasets and demonstrate that our method provides new insights into properties of GANs.} }
Endnote
%0 Conference Paper %T Geometry Score: A Method For Comparing Generative Adversarial Networks %A Valentin Khrulkov %A Ivan Oseledets %B Proceedings of the 35th International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2018 %E Jennifer Dy %E Andreas Krause %F pmlr-v80-khrulkov18a %I PMLR %P 2621--2629 %U https://proceedings.mlr.press/v80/khrulkov18a.html %V 80 %X One of the biggest challenges in the research of generative adversarial networks (GANs) is assessing the quality of generated samples and detecting various levels of mode collapse. In this work, we construct a novel measure of performance of a GAN by comparing geometrical properties of the underlying data manifold and the generated one, which provides both qualitative and quantitative means for evaluation. Our algorithm can be applied to datasets of an arbitrary nature and is not limited to visual data. We test the obtained metric on various real-life models and datasets and demonstrate that our method provides new insights into properties of GANs.
APA
Khrulkov, V. & Oseledets, I.. (2018). Geometry Score: A Method For Comparing Generative Adversarial Networks. Proceedings of the 35th International Conference on Machine Learning, in Proceedings of Machine Learning Research 80:2621-2629 Available from https://proceedings.mlr.press/v80/khrulkov18a.html.

Related Material