AI Paper Atlas

Research map / Image generation / Fast diffusion sampling

Compositional Visual Planning via Inference-Time Diffusion Scaling

Yixin Zhang, Yunhao Luo, Utkarsh Aashu Mishra, Woo Chul Shin, Yongxin Chen, Danfei Xu

ICLR 2026 · Image generation › Fast diffusion sampling · approach: Diffusion theory and design space · 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
enforce boundary agreement on Tweedie estimates instead of noisy-state score averaging (vs DiffCollage)compositional video planner · Compositional Planning Bench Overall (OOD)success rate (%): 0 -> 54Improved
use the training-free method versus trained policies on unseen start-goal pairscompositional video planner vs GCDP · Compositional Planning Bench Overall (OOD)success rate (%): 15 -> 54Improved
add asynchronous loss to synchronous message passing (ablation)message passing sampler · Cube scenesuccess rateImproved
add synchronous loss to asynchronous message passing (ablation)message passing sampler · Cube scenesuccess rateImproved
increase number of diffusion sampling steps (ablation)compositional video planner · Drawer scenesuccess rateImproved
remove discount factor in message passing (ablation)synchronous message passing · qualitative failure casesstart/goal spatial alignmentHurt
replace DiffCollage with the proposed sampler in real-world tasksreal robot Franka · Real robot Task3 (OOD)success (out of 10): 0/10 -> 10/10Improved
use proposed method rather than DiffCollage for visual qualityvideo plans · VBench++ Aesthetic (OOD)aesthetic quality: 0.46 -> 0.48No clear effect
use proposed method rather than DiffCollagevideo plans · VBench++ Imaging Quality (OOD)imaging quality: 0.55 -> 0.69Improved
use proposed method rather than DiffCollage (test-time backprop)sampling · Puzzle sampling timesampling time (s): 30.4 -> 61.8Hurt
use proposed method rather than best policy baseline on in-distribution tasksDrawer scene · Compositional Planning Bench Drawer (IND)success rate (%): 54 -> 53No clear effect

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 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