AI Paper Atlas

Research map / Video / Action recognition

CoMo: Learning Continuous Latent Motion from Internet Videos for Scalable Robot Learning

Jiange Yang, Yansong Shi, Haoyi Zhu, Mingyu Liu, Kaijing Ma, Yating Wang, Gangshan Wu, Tong He, Limin Wang

CVPR 2026 · Video › Action recognition · approach: Action anticipation and latent actions · 35 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 early temporal difference and temporal contrastive learning to continuous latent motion (ablation)CoMo · LIBERO (10 demos/task)success rate (%): 75.2 -> 80.1Improved
add early temporal difference only (ablation)continuous latent motion · LIBEROsuccess rate (%): 75.2 -> 76.9; MSE 1.6256 -> 1.5209Improved
add temporal contrastive learning only (ablation)continuous latent motion · LIBEROS-PCFC: 0.9268 -> 0.6229Improved
remove vector quantization from latent motion (naive continuous) (ablation)latent motion baseline · LIBEROS-PCFC: S-PCFC 0.9268 (near 1.0), success rate 75.2 vs discrete 75.9Hurt
use continuous instead of discrete latent motion (with Td and Tcl) (ablation)CoMo · LIBEROsuccess rate (%): 75.9 -> 80.1Improved
use continuous instead of discrete latent motion (with Td and Tcl) (ablation)auto-regressive policy · CALVIN ABC->Daverage length: 2.255 -> 3.070Improved
add Tcl to continuous baseline in auto-regressive policy (ablation)Con. · CALVIN ABC->Daverage length: 2.797 -> 2.925Improved
add Td on top of Tcl (ablation)Con.+Tcl · CALVIN ABC->Daverage length: 2.925 -> 3.070Improved
expand latent motion dimension from 128 to 512 (ablation)CoMo · CALVIN ABC->Daverage length: 3.070 -> 3.247Improved
expand latent motion dimension with Td only (no Tcl) (ablation)Con.+Td · LIBEROaction prediction MSEHurt
use future frame features as latent motion (ablation)latent motion baseline · LIBEROsuccess rate (%): MSE 2.1442 vs Dis. 5.6743, but success rate 73.8 vs 75.9Hurt
add CoMo pseudo-labelled video to diffusion policydiffusion policy · LIBEROsuccess rate (%): 70.4 -> 80.1Improved

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