Abstract
Recently, generative adversarial networks (GANs) have shown promising performance in generating realistic images. However, they often struggle in learning complex underlying modalities in a given dataset, resulting in poor-quality generated images. To mitigate this problem, we present a novel approach called mixture of experts GAN (MEGAN), an ensemble approach of multiple generator networks. Each generator network in MEGAN specializes in generating images with a particular subset of modalities, e.g., an image class. Instead of incorporating a separate step of handcrafted clustering of multiple modalities, our proposed model is trained through an end-to-end learning of multiple generators via gating networks, which is responsible for choosing the appropriate generator network for a given condition. We adopt the categorical reparameterization trick for a categorical decision to be made in selecting a generator while maintaining the flow of the gradients. We demonstrate that individual generators learn different and salient subparts of the data and achieve a multiscale structural similarity (MS-SSIM) score of 0.2470 for CelebA and a competitive unsupervised inception score of 8.33 in CIFAR-10.
| Original language | English |
|---|---|
| Title of host publication | Proceedings of the 27th International Joint Conference on Artificial Intelligence, IJCAI 2018 |
| Editors | Jerome Lang |
| Publisher | International Joint Conferences on Artificial Intelligence |
| Pages | 878-884 |
| Number of pages | 7 |
| ISBN (Electronic) | 9780999241127 |
| Publication status | Published - 2018 Jan 1 |
| Event | 27th International Joint Conference on Artificial Intelligence, IJCAI 2018 - Stockholm, Sweden Duration: 2018 Jul 13 → 2018 Jul 19 |
Publication series
| Name | IJCAI International Joint Conference on Artificial Intelligence |
|---|---|
| Volume | 2018-July |
| ISSN (Print) | 1045-0823 |
Other
| Other | 27th International Joint Conference on Artificial Intelligence, IJCAI 2018 |
|---|---|
| Country | Sweden |
| City | Stockholm |
| Period | 18/7/13 → 18/7/19 |
ASJC Scopus subject areas
- Artificial Intelligence
Cite this
Megan : Mixture of experts of generative adversarial networks for multimodal image generation. / Park, David Keetae; Yoo, Seungjoo; Bahng, Hyojin; Choo, Jaegul; Park, Noseong.
Proceedings of the 27th International Joint Conference on Artificial Intelligence, IJCAI 2018. ed. / Jerome Lang. International Joint Conferences on Artificial Intelligence, 2018. p. 878-884 (IJCAI International Joint Conference on Artificial Intelligence; Vol. 2018-July).Research output: Chapter in Book/Report/Conference proceeding › Conference contribution
}
TY - GEN
T1 - Megan
T2 - Mixture of experts of generative adversarial networks for multimodal image generation
AU - Park, David Keetae
AU - Yoo, Seungjoo
AU - Bahng, Hyojin
AU - Choo, Jaegul
AU - Park, Noseong
PY - 2018/1/1
Y1 - 2018/1/1
N2 - Recently, generative adversarial networks (GANs) have shown promising performance in generating realistic images. However, they often struggle in learning complex underlying modalities in a given dataset, resulting in poor-quality generated images. To mitigate this problem, we present a novel approach called mixture of experts GAN (MEGAN), an ensemble approach of multiple generator networks. Each generator network in MEGAN specializes in generating images with a particular subset of modalities, e.g., an image class. Instead of incorporating a separate step of handcrafted clustering of multiple modalities, our proposed model is trained through an end-to-end learning of multiple generators via gating networks, which is responsible for choosing the appropriate generator network for a given condition. We adopt the categorical reparameterization trick for a categorical decision to be made in selecting a generator while maintaining the flow of the gradients. We demonstrate that individual generators learn different and salient subparts of the data and achieve a multiscale structural similarity (MS-SSIM) score of 0.2470 for CelebA and a competitive unsupervised inception score of 8.33 in CIFAR-10.
AB - Recently, generative adversarial networks (GANs) have shown promising performance in generating realistic images. However, they often struggle in learning complex underlying modalities in a given dataset, resulting in poor-quality generated images. To mitigate this problem, we present a novel approach called mixture of experts GAN (MEGAN), an ensemble approach of multiple generator networks. Each generator network in MEGAN specializes in generating images with a particular subset of modalities, e.g., an image class. Instead of incorporating a separate step of handcrafted clustering of multiple modalities, our proposed model is trained through an end-to-end learning of multiple generators via gating networks, which is responsible for choosing the appropriate generator network for a given condition. We adopt the categorical reparameterization trick for a categorical decision to be made in selecting a generator while maintaining the flow of the gradients. We demonstrate that individual generators learn different and salient subparts of the data and achieve a multiscale structural similarity (MS-SSIM) score of 0.2470 for CelebA and a competitive unsupervised inception score of 8.33 in CIFAR-10.
UR - http://www.scopus.com/inward/record.url?scp=85055718770&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85055718770&partnerID=8YFLogxK
M3 - Conference contribution
AN - SCOPUS:85055718770
T3 - IJCAI International Joint Conference on Artificial Intelligence
SP - 878
EP - 884
BT - Proceedings of the 27th International Joint Conference on Artificial Intelligence, IJCAI 2018
A2 - Lang, Jerome
PB - International Joint Conferences on Artificial Intelligence
ER -