Self-Attention Generative Adversarial Networks
Han Zhang · Ian Goodfellow · Dimitris Metaxas · Augustus Odena

Tue Jun 11th 02:00 -- 02:20 PM @ Hall A

In this paper, we propose the Self-Attention Generative Adversarial Network (SAGAN) which allows attention-driven, long-range dependency modeling for image generation tasks. Traditional convolutional GANs generate high-resolution details as a function of only spatially local points in lower-resolution feature maps. In SAGAN, details can be generated using cues from all feature locations. Moreover, the discriminator can check that highly detailed features in distant portions of the image are consistent with each other. Furthermore, recent work has shown that generator conditioning affects GAN performance. Leveraging this insight, we apply spectral normalization to the GAN generator and find that this improves training dynamics. The proposed SAGAN performs better than prior work, boosting the best published Inception score from 36.8 to 52.52 and reducing Fr\'echet Inception distance from 27.62 to 18.65 on the challenging ImageNet dataset. Visualization of the attention layers shows that the generator leverages neighborhoods that correspond to object shapes rather than local regions of fixed shape.

Author Information

Han Zhang (Google)
Ian Goodfellow (Google Brain)
Dimitris Metaxas (Rutgers)
Augustus Odena (Google Brain)

Related Events (a corresponding poster, oral, or spotlight)

More from the Same Authors