AI Paper Atlas

Research map / Video / Video generation

MoAlign: Motion-Centric Representation Alignment for Video Diffusion Models

Aritra Bhowmik, Denis Korzhenkov, Cees G. M. Snoek, Amir Habibian, Mohsen Ghafoorian

ICLR 2026 · Video › Video generation · approach: Efficient video diffusion models · 17 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
add motion-specific (flow-supervised) features as alignment target instead of raw VideoMAEv2 features (ablation)MoAlign · VideoPhy2Joint: 23.5 -> 24.9Improved
add distance-weighted soft TRD loss instead of vanilla TRD (ablation)MoAlign · VideoPhy2Joint: 24.1 -> 24.9Improved
replace relational TRD alignment with classical REPA loss (ablation)MoAlign · VideoPhy2Joint: 24.9 -> 22.3Hurt
align at transformer layer 18 versus layer 10 (ablation)MoAlign · VideoPhy2Joint: 23.2 -> 24.9Improved
align at layer 18 versus layer 20 (ablation)MoAlign · VideoPhy2Joint: 23.8 -> 24.9Improved
MoAlign alignment instead of plain finetuning on same dataCogVideoX-2B (FT) · VideoPhy2Joint: 22.8 -> 24.9Improved
MoAlign instead of VideoREPA (reimplemented)CogVideoX-2B · VideoPhy2Joint: 23.0 -> 24.9Improved
MoAlign instead of VideoREPA on semantic adherenceCogVideoX-2B · VideoPhy2SA: 26.1 -> 28.8Improved
finetune with MoAlign on semantic adherenceCogVideoX-2B · VideoPhy (overall)SA: 49.8 -> 49.3Hurt
MoAlign instead of VideoREPA on physical commonsenseCogVideoX-2B · VideoPhy (overall)PC: 37.9 -> 39.4Improved
MoAlign alignment on VBench Total versus baseCogVideoX-2B · VBenchTotal score: 80.6 -> 81.3Improved
MoAlign alignment on VBench-2.0 Total versus baseCogVideoX-2B · VBench-2.0Total score: 54.9 -> 55.9Improved

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