AI Paper Atlas

Research map / Video / Video generation

Motion Attribution for Video Generation

Xindi Wu, Despoina Paschalidou, Jun Gao, Antonio Torralba, Laura Leal-Taixé, Olga Russakovsky, Sanja Fidler, Jonathan Lorraine

ICML 2026 · Video › Video generation · approach: Autoregressive long video generation · 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
add motion masking to attribution (ablation)Motive · VBench (Wan2.1-T2V-1.3B)Dynamic degree (%): 43.8 -> 47.6Improved
add motion masking to attribution (ablation)Motive · VBench (Wan2.2-TI2V-5B)Dynamic degree (%): 43.8 -> 48.3Improved
add motion masking to attribution (ablation)Motive · VBench (LTX-2B)Dynamic degree (%): 41.8 -> 45.1Improved
select 10% data with Motive instead of random 10%fine-tuning Wan2.1-T2V-1.3B · VBenchDynamic degree (%): 41.3 -> 47.6Improved
fine-tune on Motive 10% data instead of full datasetfine-tuning Wan2.1-T2V-1.3B · VBenchDynamic degree (%): 42.0 -> 47.6Improved
fine-tune on Motive 10% data instead of full datasetfine-tuning Wan2.2-TI2V-5B · VBenchDynamic degree (%): 45.3 -> 48.3Improved
Motive selection vs V-JEPA selectionfine-tuning Wan2.1-T2V-1.3B · VBenchDynamic degree (%): 41.6 -> 47.6Improved
Motive selection vs motion magnitude selectionfine-tuning LTX-2B · VBenchDynamic degree (%): 39.6 -> 45.1Improved
Motive selection vs full fine-tuning on imaging qualityfine-tuning Wan2.1-T2V-1.3B · VBenchImaging quality (%): 63.9 -> 64.6Hurt
add Motive-selected fine-tuning vs base model (human preference)Wan2.1-T2V-1.3B · Human evaluationWin rate (%): 74.1Improved
add motion masking (human preference) (ablation)Motive · Human evaluationWin rate (%): 46.9 win, 33.1 lossImproved
increase projection dimension from 128 to 512 (ablation)Motive gradient projection · Gradient ranking preservationSpearman correlation (%): 46.9 -> 74.7Improved

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