AI Paper Atlas

Research map / Image generation / GANs

Unifying GANs and Score-Based Diffusion as Generative Particle Models

Jean-Yves Franceschi, Mike Gartrell, Ludovic Dos Santos, Thibaut Issenhuth, Emmanuel de Bezenac, Mickael Chen, Alain Rakotomamonjy

NeurIPS 2023 · Image generation › GANs · approach: Stabilized GAN training · 2 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
Replace EDM (generator-free diffusion) with Discriminator Flowparticle generative model · MNISTFID: 3 -> 4No clear effect
Replace EDM with Discriminator Flowparticle generative model · CelebA 64x64FID: 10 -> 41Hurt
Replace GAN with Score GAN (score-based generator training)generator-based model · MNISTFID: 3 -> 15Hurt
Replace GAN with Score GAN (score-based generator training)generator-based model · CelebA 64x64FID: 19 -> 35Hurt
Add a generator to diffusion (Score GAN instead of EDM) for one-step samplingEDM · MNISTsampling cost / FID: 3 -> 15Mixed
Remove generator from GAN (Discriminator Flow instead of GAN) on MNISTGAN · MNISTFID: 3 -> 4No clear effect
Remove generator from GAN (Discriminator Flow instead of GAN) on CelebAGAN · CelebA 64x64FID: 19 -> 41Hurt
Use Discriminator Flow instead of EDM for early stopping of generationEDM · MNISTconvergence speed to data distributionImproved
Use Discriminator Flow with fewer steps vs EDM Euler samplerDiscriminator Flow · MNISTFID vs NFEHurt
Vary number of score update steps K from 1 to 10 in Score GAN (ablation)Score GAN · MNISTFIDNo clear effect
Add stochastic component (Brownian noise) to Score GAN updates (ablation)Score GAN · image dataFIDHurt
Use EDM parameters (mu_data=2, mu_rho=1) in Score GAN instead of NCSN parameters (ablation)Score GAN · MNISTmode collapseHurt

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