AI Paper Atlas

Research map / Image generation / Fast diffusion sampling

Strictly Constrained Generative Modeling via Split Augmented Langevin Sampling

Matthieu Blanke, Yongquan Qu, Sara Shamekh, Pierre Gentine

ICLR 2026 · Image generation › Fast diffusion sampling · approach: Distributional control of diffusion · 7 citations (OpenAlex)

Read the paper

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
vary rho over iterations (linear 2 to 20) instead of constant rho_min (ablation)CASAL · Field generation, 1000 stepswrong-mode proportion: 42.4% -> 0.04%Improved
vary rho linearly instead of constant rho_max (ablation)CASAL · Field generation, 1000 stepswrong-mode proportion: 22% -> 0.04%Improved
use linear schedule instead of logarithmic (ablation)CASAL · Field generation, 1000 stepswrong-mode proportion: 3.8% -> 0.04%Improved
increase Langevin steps from 1000 to 5000 with linear schedule (ablation)CASAL · Field generationwrong-mode proportion: 0.04% -> 0.001%Improved
add CASAL splitting vs projected diffusionprojected diffusion · Darcy flowreconstruction error: 22% -> 10%Improved
use CASAL instead of penalty guidancepenalty diffusion · Darcy flowconstraint violation: 25% -> 2e-5%Improved
use CASAL instead of projected diffusionprojected diffusion · Navier-Stokesreconstruction error: 10% -> 8%Improved
use CASAL instead of projected diffusionprojected diffusion · Poisson equationreconstruction error: 15% -> 5%Improved
use CASAL instead of projected Langevinprojected Langevin · Data assimilation runtimeseconds per 10^6 steps: 2.46 -> 0.31Improved
use CASAL instead of unconstrained Langevinunconstrained Langevin · Fluid generation runtimeseconds per 10^6 steps: 0.37 -> 0.45Hurt
use CASAL instead of projected Langevinprojected Langevin · Energy-constrained field generationmatch to conditional distributionImproved
use CASAL instead of penalty guidancepenalty guidance · Quadrotor feasibilitysuccess rateImproved

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