Generative Adversarial Networks: From Minimax to StyleGAN
Design a realistic face generator — the two-player minimax game, the mode-collapse/vanishing-gradient/instability ladder and their named fixes (WGAN-GP, progressive growing), StyleGAN's mapping network and style injection, latent-space editing, and the GAN-vs-diffusion-vs-autoregressive comparison the rest of the generative-vision track builds on
The foundation subject the rest of this track's business cases depend on: why a single forward pass still matters in 2026 (real-time serving, the adversarial loss hiding inside every VAE/VQGAN decoder, GAN-based distillation of diffusion models, still-SOTA perceptual super-resolution and fast inpainting) argued against diffusion and autoregressive generation on latency, quality, diversity, training stability and controllability; the generator-vs-discriminator minimax game and the non-saturating loss at interview altitude; the failure-mode ladder — mode collapse, vanishing gradients, high-resolution instability — and each fix (minibatch discrimination, WGAN/WGAN-GP, progressive growing); StyleGAN's mapping network, AdaIN/modulated convolutions, per-layer style injection, noise inputs, style mixing and the truncation trick, with StyleGAN2's fixes at one-paragraph altitude; controllability as a product feature (InterFaceGAN-style latent directions, GAN inversion, identity preservation); FFHQ-style data and the output-diversity requirement; FID/IS plus precision/recall for generative models and a bias audit; and a synchronous generation-service design with latent storage, moderation and the consent/deepfake question.
Practice questions (6)
-
View →
Real-Time Avatars: Arguing GAN Over Diffusion
Advanced · Free -
View →
Debugging a Generator That Stopped Improving
Advanced · Free -
View →
A Discriminator That's Too Good, Too Fast
Advanced -
View →
Designing an Attribute-Editing Feature on StyleGAN
Advanced -
View →
A Suspiciously Good FID Score
Advanced