SAN: Inducing Metrizability of GAN with Discriminative Normalized Linear Layer
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2023
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866916198719946752 |
|---|---|
| author | Takida, Yuhta Imaizumi, Masaaki Shibuya, Takashi Lai, Chieh-Hsin Uesaka, Toshimitsu Murata, Naoki Mitsufuji, Yuki |
| author_facet | Takida, Yuhta Imaizumi, Masaaki Shibuya, Takashi Lai, Chieh-Hsin Uesaka, Toshimitsu Murata, Naoki Mitsufuji, Yuki |
| contents | Generative adversarial networks (GANs) learn a target probability distribution by optimizing a generator and a discriminator with minimax objectives. This paper addresses the question of whether such optimization actually provides the generator with gradients that make its distribution close to the target distribution. We derive metrizable conditions, sufficient conditions for the discriminator to serve as the distance between the distributions by connecting the GAN formulation with the concept of sliced optimal transport. Furthermore, by leveraging these theoretical results, we propose a novel GAN training scheme, called slicing adversarial network (SAN). With only simple modifications, a broad class of existing GANs can be converted to SANs. Experiments on synthetic and image datasets support our theoretical results and the SAN's effectiveness as compared to usual GANs. Furthermore, we also apply SAN to StyleGAN-XL, which leads to state-of-the-art FID score amongst GANs for class conditional generation on ImageNet 256$\times$256. Our implementation is available on https://ytakida.github.io/san. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2301_12811 |
| institution | arXiv |
| publishDate | 2023 |
| record_format | arxiv |
| spellingShingle | SAN: Inducing Metrizability of GAN with Discriminative Normalized Linear Layer Takida, Yuhta Imaizumi, Masaaki Shibuya, Takashi Lai, Chieh-Hsin Uesaka, Toshimitsu Murata, Naoki Mitsufuji, Yuki Machine Learning Generative adversarial networks (GANs) learn a target probability distribution by optimizing a generator and a discriminator with minimax objectives. This paper addresses the question of whether such optimization actually provides the generator with gradients that make its distribution close to the target distribution. We derive metrizable conditions, sufficient conditions for the discriminator to serve as the distance between the distributions by connecting the GAN formulation with the concept of sliced optimal transport. Furthermore, by leveraging these theoretical results, we propose a novel GAN training scheme, called slicing adversarial network (SAN). With only simple modifications, a broad class of existing GANs can be converted to SANs. Experiments on synthetic and image datasets support our theoretical results and the SAN's effectiveness as compared to usual GANs. Furthermore, we also apply SAN to StyleGAN-XL, which leads to state-of-the-art FID score amongst GANs for class conditional generation on ImageNet 256$\times$256. Our implementation is available on https://ytakida.github.io/san. |
| title | SAN: Inducing Metrizability of GAN with Discriminative Normalized Linear Layer |
| topic | Machine Learning |
| url | https://arxiv.org/abs/2301.12811 |