AI Paper Atlas

Research map / Video / Action recognition

DreamDojo: A Generalist Robot World Model from Large-Scale Human Videos

Shenyuan Gao, William Liang, Kaiyuan Zheng, Ayaan Naveed Malik, Seonghyeon Ye, Sihyun Yu, Wei-Cheng Tseng, Yuzhu Dong, Kaichun Mo, Chen-Hsuan Lin, Jiannan Xiang, Yuqi Xie and 11 more

ICML 2026 · Video › Action recognition · approach: Action anticipation and latent actions · 1 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
DreamDojo (ablation)In-lab Eval · pretrain with continuous latent actions instead of action-free videoPSNR: 20.797 -> 20.913Improved
DreamDojo (ablation)EgoDex Eval · pretrain with continuous latent actions instead of action-free videoPSNR: 19.924 -> 20.344Improved
Cosmos-Predict2.5 (ablation)In-lab / EgoDex Eval · action-free pretraining on human videos vs no pretrainingPSNR: 20.576 -> 20.797 (In-lab); 19.952 -> 19.924 (EgoDex)No clear effect
DreamDojo (ablation)In-lab Eval · use ground-truth actions instead of latent actionsPSNR: 20.913 -> 20.960Improved
DreamDojo (ablation)In-lab Eval · add EgoDex to In-lab pretraining mixturePSNR: 20.913 -> 20.972Improved
DreamDojo (ablation)DreamDojo-HV Eval · add DreamDojo-HV to pretraining mixturePSNR: 18.706 -> 18.724Improved
DreamDojoCounterfactual Eval · scale model from 2B to 14BPSNR: 20.907 -> 21.087Improved
Cosmos-Predict2.5 (ablation)Counterfactual Eval · pretrain on human videos vs nonePSNR: 20.472 -> 20.907Improved
DreamDojo-2BGR-1 Long Eval · distill teacher into few-step autoregressive studentPSNR: 14.086 -> 13.146Mixed
DreamDojo-2BGR-1 Long Eval · distill teacher into few-step autoregressive studentFPS: 2.72 -> 10.93Improved
GR00T N1.5 policyAgiBot fruit packing · apply model-based planning with DreamDojo vs uniform samplingsuccess rate: 55.33% -> 68.00%Improved

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