AI Paper Atlas

Research map / Video / Video generation

DreamRunner: Fine-Grained Compositional Story-to-Video Generation with Retrieval-Augmented Motion Adaptation

Zun Wang, Jialu Li, Han Lin, Jaehong Yoon, Mohit Bansal

AAAI 2026 · Video › Video generation · approach: Compositional text-to-video diffusion · 17 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
add SR3AI region-based attention and LoRA injection (ablation)DreamRunner (CogVideoX-2B) · DreamStorySet ablationtransition DINO: 87.1 -> 92.5Improved
add retrieval-augmented motion priors (ablation)DreamRunner (CogVideoX-2B) · DreamStorySet ablationfine-grained text ViCLIP: 22.5 -> 23.1Improved
add retrieval-augmented motion priors without SR3AI (ablation)CogVideoX-2B · DreamStorySet ablationtransition DINO: 87.1 -> 84.6Hurt
add RAG on top of SR3AI (ablation)DreamRunner · DreamStorySet ablationtransition DINO: 92.5 -> 93.6Mixed
use per-video prompts instead of a single prompt for RAG motion training (ablation)CogVideoX-2B + RAG · 64-motion promptsViCLIP: 22.02 -> 23.04Improved
add RAG motion prior learning (ablation)CogVideoX-2B · 64-motion promptsViCLIP: 20.84 -> 23.04Improved
add SR3A to CogVideoX-5BCogVideoX-5B · T2V-CompBenchDynamic-attr: 0.2250 -> 0.2832Improved
add SR3A to CogVideoX-5BCogVideoX-5B · T2V-CompBenchSpatial: 0.5845 -> 0.6875Improved
add SR3A to CogVideoX-5BCogVideoX-5B · T2V-CompBenchAction: 0.6040 -> 0.6340Improved
DreamRunner vs VLogger on character consistencyDreamRunner · DreamStorySetcharacter DINO: 41.3 -> 55.1Improved
DreamRunner vs VLogger on video smoothnessDreamRunner · DreamStorySetvideo smoothness: 96.2 -> 98.1Improved
region-based LoRA injection vs global LoRA injection (ablation)CogVideoX-2B · qualitative multi-charactervisual qualityImproved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 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