AGAN: Towards Automated Design of Generative Adversarial Networks

By Hanchao Wang, Jun Huan. 2019

41 citations recorded by the source corpus.

Recent progress in Generative Adversarial Networks (GANs) has shown promising signs of improving GAN training via architectural change. Despite some early success, at present the design of GAN architectures requires human expertise, laborious trial-and-error testings, and often draws inspiration from its image classification counterpart. In the current paper, we present the first neural architecture search algorithm, automated neural architecture search for deep generative models, or AGAN for abbreviation, that is specifically suited for GAN training. For unsupervised image generation tasks on CIFAR-10, our algorithm finds architecture that outperforms state-of-the-art models under same regularization techniques. For supervised tasks, the automatically searched architectures also achieve highly competitive performance, outperforming best human-invented architectures at resolution $32\times32$. Moreover, we empirically demonstrate that the modules learned by AGAN are transferable to other image generation tasks such as STL-10.

Subfields: Computer Vision and Pattern Recognition

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Vision and Pattern Recognition

1. Shuicheng Yan
2. Si Liu
3. Yunpeng Chen
4. Chen Gao
5. Jiashi Feng
6. Dacheng Tao
7. Ming-Hsuan Yang
8. Luc Van Gool

Citations in the corpus

• AdversarialNAS: Adversarial Neural Architecture Search for GANs (2019, methodology): “Following the setting of AutoGAN [12] and AGAN [41], we directly apply the generator searched on CIFAR-10 to STL-10 dataset for evaluating the transferability of the architecture.”