AI Paper Atlas

Research map / Video / Video generation

Retrieving and Refining Winning Noise Tickets for Diffusion-Based Motion Generation

Sakuya Ota, Qing Yu, Kent Fujiwara, Satoshi Ikehata, Ikuro Sato

ECCV 2026 · Video › Video generation · approach: Text-to-motion diffusion · 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
retrieve semantically aligned noise from null-prompt dictionary (WIN-R) (ablation)MDM-50step · HumanML3DFID: 0.438 -> 0.250Improved
add frame-adjusted retrieval with perturbation (WIN-FAR) over plain retrieval (ablation)WIN-R · HumanML3D MDM-50stepR-Top1: 0.505 -> 0.512Improved
add KL-regularized noise refinement after retrieval (WINRO) (ablation)WIN-FAR · HumanML3D MDM-50stepFID: 0.248 -> 0.115Improved
add KL-regularized refinement (ablation)WIN-FAR · HumanML3D MotionLCMR-Top1: 0.572 -> 0.580Improved
add LoRA noise refiner instead of iterative optimization (ablation)WINRO · HumanML3D MDM-50stepAITS (s): 81.408 -> 0.280Improved
add KL regularization term (lambda=250 vs 0) (ablation)WINRO · HumanML3D MotionLCMFID: 0.136 -> 0.066Improved
use cosine similarity instead of Euclidean distance (ablation)WINRO · HumanML3D MotionLCMFID: 0.095 -> 0.066Improved
increase dictionary size from 100 to 10k (ablation)WIN-FAR · HumanML3D MotionLCMFID: 0.094 -> 0.083Improved
increase dictionary size beyond 10k to 100k (ablation)WIN-FAR · HumanML3D MotionLCMFID: 0.083 -> 0.085No clear effect
vary LoRA rank from 16 to 128 (ablation)WIN-LoRA · HumanML3D MotionLCMR-Top1: 0.553-0.560No clear effect
apply WINRO to MTT single-text generation (ablation)MDM-50step · MTTR@1: 12.2 -> 20.7Improved
apply optimization (WIN-O) to STMC multi-track long generation (ablation)STMC · MTTR@1: 33.6 -> 62.1Improved

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