Our basic setup follows SA-GAN (Zhang et al., 2018), and is implemented in TensorFlow (Abadi et al., 2016). We employ the architectures detailed in Appendix B, with non-local blocks inserted at a single stage in each network. Both G and D n…This leads to two conclusions: first, as has been noted in previous works (Miyato et al., 2018; Gulrajani et al., 2017; Zhang et al., 2018), D must remain optimal with respect to G both for stability and to provide useful gradient informati…
Miyato et al.
other · 2 mentions across 1 reading
In this course
Miyato et al. are cited for their work on stabilizing discriminator training in generative adversarial networks, specifically addressing the technical requirement that the discriminator remain optimal relative to the generator. Their contribution appears foundational to understanding GAN stability—a core concern when deploying these models for generative tasks in art and media contexts that the course examines.
Mentioned in 1 reading
Appears alongside
People mentioned in the same passages — sorted by co-occurrence weight.