AI Paper Atlas

Research map / Image generation / GANs

Toward a Visual Concept Vocabulary for GAN Latent Space

Sarah Schwettmann, Evan Hernandez, David Bau, Samuel Klein, Jacob Andreas, Antonio Torralba

ICCV 2021 · Image generation › GANs · approach: Large-scale GAN latent analysis · 12 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
use layer-selective directions instead of random directions (ablation)annotation direction generation · BigGAN-Places direction annotationsdistinct 1-grams: 2,316 -> 3,156Improved
use layer-selective directions instead of GANSpace directions (ablation)annotation direction generation · BigGAN-Places direction annotationsdistinct 1-grams: 2,975 -> 3,156Improved
use layer-selective directions instead of random directions (ablation)annotation direction generation · BigGAN-Places direction annotationsdistinct 3-grams: 22,938 -> 31,307Improved
use layer-selective directions instead of GANSpace directions (ablation)annotation direction generation · BigGAN-Places direction annotationsinter-annotator BERTScore-R: 0.343 -> 0.393Improved
use layer-selective directions instead of random directions (ablation)annotation direction generation · BigGAN-Places direction annotationsinter-annotator BLEU: 8.86 -> 7.17Hurt
apply distilled directions to new z (same class)distilled concept directions · BigGAN-Placeshuman accuracy (4-way): more than 60%Improved
apply distilled directions to unseen classes instead of the same classdistilled concept directions · BigGAN-Placeshuman accuracy (4-way): .66 -> .39Hurt
compose pairs of distilled directions instead of single onesdistilled concept directions · BigGAN-Placeshuman accuracy (4-way): .66 -> .44Hurt

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy