SkeLAM splits robot latent actions into skeleton and residual
On DROID, end-effector ADE falls 19.5% and LPIPS 11.1% versus DreamDojo, with zero-shot human-to-robot transfer shown qualitatively.
Stories per day · 30 days
Relatedrobotics3world-models23d1depth1hierarchical-planning1jepa1latent-actions1manipulation1skelam1skeleton1stable-video-diffusion1visual-planning1world-model1
3 stories, newest first · one card per story
Stories tagged #droid
DepthWorld: 3D World Model for Robot Manipulation
DepthWorld jointly predicts multi-view RGB and metric depth for robot manipulation from calibrated 3D data.
H-JEPA: End-to-End Learning of Hierarchical World Models for Visual Planning
H-JEPA trains hierarchical action-conditioned world models, raising Visual AntMaze planning success from 18% to 73%.