AI Paper Atlas

Research map / Image generation / Fast diffusion sampling

Polarity Sampling: Quality and Diversity Control of Pre-Trained Generative Networks via Singular Values

Ahmed Imtiaz Humayun, Randall Balestriero, Richard Baraniuk

CVPR 2022 · Image generation › Fast diffusion sampling · approach: Distributional control of diffusion · 19 citations (OpenAlex)

Check 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
Apply polarity sampling at ρ=0.2StyleGAN2 · LSUN Church 256×256FID: 6.29 -> 6.02Improved
Apply polarity sampling in pixel space at ρ=-0.001StyleGAN2 · LSUN Church 256×256FID: 3.97 -> 3.92Improved
Apply polarity sampling in style space at ρ=0.01StyleGAN2 · LSUN Car 512×384FID: 2.34 -> 2.27Improved
Apply polarity sampling in style space at ρ=-0.1StyleGAN2 · LSUN Cat 256×256FID: 6.49 -> 6.39Improved
Apply polarity sampling in InceptionV3 space at ρ=0.01StyleGAN2-F · FFHQ 1024×1024FID: 2.74 -> 2.57Improved
Apply polarity sampling in VGG space at ρ=-0.001StyleGAN3-T · AFHQv2 512×512FID: 4.05 -> 3.95Improved
Apply pixel-space polarity sampling at ρ=-0.01ProGAN · CelebAHQ 1024×1024FID: 7.37 -> 7.28Improved
Increase the number of latent samples from N=100K to N=500K (ablation)StyleGAN2 · FFHQ 1024×1024FID (lowest): 2.63 -> 2.58Improved
Increase the number of latent samples from N=100K to N=500K (ablation)StyleGAN2 · LSUN Cat 256×256FID (lowest): 6.38 -> 6.34No clear effect
Use K=40 top singular values rather than K=10 (ablation)StyleGAN2 · FFHQ 1024×1024FID (lowest): 2.71 -> 2.57Improved
Use K=100 rather than K=40 top singular values (ablation)StyleGAN2 · FFHQ 1024×1024FID (lowest): 2.57 -> 2.70Hurt
Use positive polarity to increase diversityStyleGAN2 with ψ<1 · FFHQFID: {.69, 8.11, 11.1} pointsImproved

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