AI Paper Atlas

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

Source-Guided Flow Matching

Zifan Wang, Alice Harting, Matthieu Barreau, Michael M. Zavlanos, Karl Henrik Johansson

ICLR 2026 · Probabilistic and scientific modeling › Variational and generative sampling · approach: Flow matching and flow sampling

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
modify source distribution (SGFM) instead of the vector field, with exact transport mapSGFM · theorydistribution recoveryImproved
use chi-square norm regularizer (D-Flow) instead of plain Gaussian prior regularizer in optimization sampling (ablation)SGFM-OPT · Darcy flow validityvalidity: 0.907 -> 0.474Improved
use HMC sampling instead of plain MAP optimization (ablation)SGFM · Darcy flow validityvalidity: 0.907 -> 0.591Improved
SGFM-OPT chi^2 vs vector-field guidance g-cov-ASGFM · Darcy flow validityvalidity: 0.993 -> 0.474Improved
use SGFM-OPT-4 (shell regularizer) vs g-covA/g-covGSGFM · CelebA denoisingPSNR: 30.35 -> 31.60Improved
SGFM-OPT-4 vs PnP-flow on denoisingSGFM · CelebA denoisingPSNR: 32.14 -> 31.60Hurt
SGFM-OPT-4 vs PnP-flow on deblurringSGFM · CelebA deblurringPSNR: 38.74 -> 35.27Hurt
SGFM-OPT-5 vs PnP-flow on super-resolutionSGFM · CelebA super-resolutionPSNR: 31.33 -> 33.33Improved
use projected gradient shell constraint (SGFM-6) instead of regularizer variants (ablation)SGFM-OPT · CelebA deblurringPSNR: 35.27 -> 32.60Hurt
use optimization-based sampling instead of exact sampling on S-shaped exampleSGFM / D-Flow · 2D S-shapediversityHurt
increase importance-sampling particle count (ablation)SGFM-IS · 2D toyWasserstein errorImproved
reduce NFEs with optimal vector field (ablation)SGFM-IS · 2D toysample qualityNo clear effect

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