Adversarial Loss

Concept

A strategy using a discriminator network to distinguish between real and generated images, penalizing the generator for blurry or fake-looking outputs to improve realism.

Mentioned in 1 video

Videos Mentioning Adversarial Loss

Stanford CME296 Diffusion & Large Vision Models | Spring 2026 | Lecture 4 - Latent Space & Guidance

Stanford CME296 Diffusion & Large Vision Models | Spring 2026 | Lecture 4 - Latent Space & Guidance

Stanford Online

A strategy using a discriminator network to distinguish between real and generated images, penalizing the generator for blurry or fake-looking outputs to improve realism.