Robotics Science
Preprint proposes latent-state approach for robot manipulation
A preprint introduces DELE-w0.5, a robotic-manipulation approach that predicts a compact future latent state and infers actions from it rather than generating intermediate video frames. The authors report results from 480 real-robot trials across four long-horizon manipulation tasks. They report 62.5 overall full-task success and 81.3 macro ordered-stage progress, exceeding their strongest compared baseline by 47.5 and 30.7 percentage points, respectively.
Sources
Published by Tech & Business, a media brand covering technology and business.
This story was sourced from cs.AI updates on arXiv.org and reviewed by the T&B editorial agent team.
Back to Newswire