Presentation | 2018-11-05 [Poster Presentation] How does the complexity of critic affect the performance of WGAN? Akihiro Iohara, Toshiyuki Tanaka, |
---|---|
PDF Download Page | PDF download Page Link |
Abstract(in Japanese) | (See Japanese page) |
Abstract(in English) | WGAN is a generative model that learns by minimizing the Wasserstein distance between the generator distribution and the real-data distribution, evaluated via the Monge-Kantorovich dual formulation, a maximization problem with respect to the critic. Since the critic is typically implemented as a neural network, its complexity may affect accuracy of evaluated Wasserstein distances, and consequently, performance of WGAN. In this paper, we exper- imentally study the relationship between complexity of the critic and performance of WGAN by using the empirical Wasserstein distance as well as other GAN evaluation metrics. |
Keyword(in Japanese) | (See Japanese page) |
Keyword(in English) | Generative modeling / GANs / Wasserstein GAN / Optimal transport |
Paper # | IBISML2018-60 |
Date of Issue | 2018-10-29 (IBISML) |
Conference Information | |
Committee | IBISML |
---|---|
Conference Date | 2018/11/5(3days) |
Place (in Japanese) | (See Japanese page) |
Place (in English) | Hokkaido Citizens Activites Center (Kaderu 2.7) |
Topics (in Japanese) | (See Japanese page) |
Topics (in English) | Information-Based Induction Science Workshop (IBIS2018) |
Chair | Hisashi Kashima(Kyoto Univ.) |
Vice Chair | Masashi Sugiyama(Univ. of Tokyo) / Koji Tsuda(Univ. of Tokyo) |
Secretary | Masashi Sugiyama(Nagoya Inst. of Tech.) / Koji Tsuda(AIST) |
Assistant | Tomoharu Iwata(NTT) / Shigeyuki Oba(Kyoto Univ.) |
Paper Information | |
Registration To | Technical Committee on Infomation-Based Induction Sciences and Machine Learning |
---|---|
Language | JPN |
Title (in Japanese) | (See Japanese page) |
Sub Title (in Japanese) | (See Japanese page) |
Title (in English) | [Poster Presentation] How does the complexity of critic affect the performance of WGAN? |
Sub Title (in English) | |
Keyword(1) | Generative modeling |
Keyword(2) | GANs |
Keyword(3) | Wasserstein GAN |
Keyword(4) | Optimal transport |
1st Author's Name | Akihiro Iohara |
1st Author's Affiliation | Kyoto University(Kyoto Univ.) |
2nd Author's Name | Toshiyuki Tanaka |
2nd Author's Affiliation | Kyoto University(Kyoto Univ.) |
Date | 2018-11-05 |
Paper # | IBISML2018-60 |
Volume (vol) | vol.118 |
Number (no) | IBISML-284 |
Page | pp.pp.119-125(IBISML), |
#Pages | 7 |
Date of Issue | 2018-10-29 (IBISML) |