AI Paper Atlas

Research map / Probabilistic and scientific modeling / Variational and generative sampling

Dimension-Free Multimodal Sampling via Preconditioned Annealed Langevin Dynamics

Lorenzo Baldassari, Josselin Garnier, Knut Solna, Maarten V. de Hoop

ICML 2026 · Probabilistic and scientific modeling › Variational and generative sampling · approach: Posterior sampling with diffusion guidance · 0 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 decaying smoothing and preconditioner spectra (Gamma=Diag(j^-1.5), C=Diag(40 j^-2.7)) instead of flat Gamma=I, C=40I (ablation)annealed Langevin dynamics · 2-component Gaussian mixture, exact score, dimension sweepannealing-induced KLImproved
use decaying smoothing and preconditioner spectra instead of flat spectra (ablation)annealed Langevin dynamics · bimodal Gaussian mixture, steps to reach KL 0.3ALD steps to reach accuracyImproved
use a fast-decaying preconditioner Gamma=Diag(j^-3.5) instead of Diag(j^-1) under covariance-perturbed score (ablation)annealed Langevin dynamics · bimodal Gaussian mixture with misspecified score, dimension sweepempirical KL vs dimensionImproved
increase preconditioner scale to reduce annealing bias boundannealed Langevin dynamics · theory, Euler-Maruyama discretizationstabilityMixed

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