AI Paper Atlas

Research map / Reinforcement learning / Visual and representation RL

World Models for Learning Dexterous Hand-Object Interactions from Human Videos

Raktim Goswami, Amir Bar, David Fan, Tsung-Yen Yang, Gaoyue Zhou, Prashanth Krishnamurthy, Michael Rabbat, Farshad Khorrami, Yann LeCun

ECCV 2026 · Reinforcement learning › Visual and representation RL · approach: Dexterous manipulation transfer · 48 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 hand consistency loss (ablation)DexWM · EgoDex open-loopPCK@20 at 4s: 26 -> 60Improved
add hand consistency loss (ablation)DexWM · EgoDex open-loopembedding L2 error at 4s: 0.85 -> 0.66Improved
add EgoDex human video on top of DROID (ablation)DexWM · RoboCasa Lift open-loopPCK@20 at 4s: 2 -> 7Improved
add DROID to EgoDex training (ablation)DexWM · EgoDex open-loopPCK@20 at 4s: 60 -> 60No clear effect
human-video pretraining before robot fine-tuning (ablation)DexWM · RoboCasa graspsuccess rate (%): 14 -> 58Improved
human-video pretraining before robot fine-tuning (ablation)DexWM · RoboCasa reachsuccess rate (%): 18 -> 72Improved
pretrain Diffusion Policy on the same human data (ablation)Diffusion Policy · RoboCasa avgsuccess rate (%): 4% vs DexWM 53%No clear effect
use world model with MPC instead of goal-conditioned diffusion policyDexWM · real robot graspsuccess rate (%): 0 -> 83Improved
increase predictor size S to XL (31M to 456M) (ablation)DexWM · open-loop EgoDex/RoboCasaembedding error and PCK@20Improved
swap DINOv2 for DINOv3, Web-SSL, V-JEPA 2 or SigLIP 2 (ablation)DexWM · RoboCasa sim taskssuccess rateMixed
add keypoint cost to latent state planning cost (ablation)DexWM · robot planningsuccess rateImproved
Lift-initialized exploration instead of random-target exploration (ablation)DexWM · RoboCasa avgsuccess rate (%): 49 -> 53Improved

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