AI Paper Atlas

Research map / Image generation / Fast diffusion sampling

Trajectory Stitching for Solving Inverse Problems with Flow-Based Models

Alexander Denker, Zeljko Kereta, Carola-Bibiane Schönlieb, Moshe Eliasof

ICML 2026 · Image generation › Fast diffusion sampling · approach: Rectified flow straightening · 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
replace single-shooting D-Flow with multiple-shooting MS-FlowD-Flow · CelebA Gaussian deblurringPSNR (dB): 36.07 -> 39.04Improved
replace D-Flow with MS-FlowD-Flow · CelebA 2x super-resolutionPSNR (dB): 35.49 -> 36.41Improved
replace D-Flow with MS-FlowD-Flow · CelebA 4x super-resolutionPSNR (dB): 29.93 -> 28.52Hurt
replace D-Flow with MS-FlowD-Flow · CelebA box-inpaintingPSNR (dB): 34.20 -> 34.77Mixed
replace D-Flow with MS-FlowD-Flow · CelebA Gaussian deblurringSSIM: 0.957 -> 0.968Mixed
replace D-Flow with MS-FlowD-Flow · CelebA image recoverytime per image (s): 26.31 -> 14.11Improved
use MS-Flow with JFB (L=1) instead of D-Flow at n_t=12D-Flow · OrganCMNIST CTcomputation time (s): 175.57 -> 4.44Improved
add Jacobian-free approximation to MS-Flow (L=1, n_t=12) (ablation)MS-Flow w/o JFB · OrganCMNIST CTcomputation time (s): 41.05 -> 4.44Improved
increase inner iterations L from 1 to 10 (n_t=12) (ablation)MS-Flow w/ JFB · OrganCMNIST CTcomputation time (s): 4.44 -> 40.10Hurt
use exact gradients instead of Jacobian-free approximation (ablation)MS-Flow coordinate descent · OrganCMNIST CTtrajectory loss convergenceNo clear effect
replace FlowDPS with MS-FlowFlowDPS · FFHQ deblurring noise 0.01PSNR (dB): 27.78 -> 30.65Improved
replace FLAIR with MS-FlowFLAIR · FFHQ deblurring noise 0.01LPIPS: 0.199 -> 0.268Hurt

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 3 of 3 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

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