AI Paper Atlas

Research map / Video / Video generation

UMO: Unified In-Context Learning Unlocks Motion Foundation Model Priors

Xiaoyan Cong, Zekun Li, Zhiyang Dou, Hongyu Li, Omid Taheri, chuan guo, Abhay Mittal, Sizhe An, Taku Komura, Wojciech Matusik, Michael Black, Srinath Sridhar

ECCV 2026 · Video › Video generation · approach: Text-to-motion diffusion · 7 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
use temporal fusion instead of AdaLN injection (ablation)UMO · Keyframe infilling[P]-MPJPE: 11.1 -> 0.95Improved
use temporal fusion instead of sequential concatenation (ablation)UMO · Keyframe infillingFID: 11.77 -> 0.476Improved
use temporal fusion instead of ControlNet (ablation)UMO · Keyframe infilling[P]-MPJPE: 5.19 -> 0.95Improved
use AdaLN instead of temporal fusion (ablation)UMO · Keyframe infillingR@1: 0.692 -> 0.746Mixed
use temporal fusion instead of sequential concatenation (ablation)UMO · Keyframe infillinglatency overhead (s): 0.89 -> 0.01Improved
add [edit] meta-operation embedding to the two-token set (ablation)UMO-Unified · Temporal inpainting (keyframe infilling)[P]-MPJPE: 1.18 -> 0.95Improved
add [edit] meta-operation embedding to the two-token set (ablation)UMO-Unified · MotionFix editing (full)AvgR: 1.97 -> 1.75Improved
add [edit] meta-operation embedding to the two-token set (ablation)UMO-Unified · Temporal inpainting (prediction)[P]-MPJPE: 0.71 -> 0.54Improved
joint multi-task training (Unified) instead of single-task ExpertUMO · HumanML3D text-to-motionFID: 17.04 -> 9.460Improved
joint multi-task training (Unified) instead of single-task ExpertUMO · Temporal inpainting (in-betweening)FID: 0.340 -> 0.050Improved
fine-tune HY-Motion with UMO instead of using it zero-shotHY-Motion · HumanML3D text-to-motionFID: 61.035 -> 17.04Improved
fine-tune HY-Motion with UMO instead of training-free inversionHY-Motion · Temporal inpainting (keyframe infilling)FID: 36.82 -> 0.075Improved

Earlier methods this paper compares with

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