robotics
Ten Thousand Hours of Video Became a Robot Simulator
WorldLine separated transferable visual dynamics from action grounding across incompatible robot bodies.
Summary
WorldLine separated transferable visual dynamics from action grounding across incompatible robot bodies.
WorldLine learns manipulation dynamics from more than 10,000 hours of action-free robot video, then grounds those dynamics with over 2,000 hours of action trajectories spanning more than ten embodiments. An image-space action representation provides a shared control interface, while failure-enriched and multi-view training target interaction fidelity. The authors report 74% mean trajectory-success prediction across RoboTwin and AgiBot, and gains of up to 21.4 percentage points when its rollouts guide tasks without RoboTwin training or adaptation. Those are benchmark results, not a substitute for hardware safety checks.
Why it matters
WorldLine separated transferable visual dynamics from action grounding across incompatible robot bodies.
Limits and context
- Those are benchmark results, not a substitute for hardware safety checks.
Key claims
WorldLine separated transferable visual dynamics from action grounding across incompatible robot bodies.
Qualification: Those are benchmark results, not a substitute for hardware safety checks.
Evidence: source-2026-09-30-013
Sources
- arXiv preprint 2609.38059arXiv · primary research
Corrections
No corrections have been recorded for this story.