robotics
Alignment Beat Raw Hours in Human-to-Robot Data
Ego4WAM separated duration, diversity, action labels and human-robot alignment under one world-action model.
Summary
Ego4WAM separated duration, diversity, action labels and human-robot alignment under one world-action model.
Ego4WAM holds the model backbone fixed while varying the properties and uses of egocentric human video. In real-robot and RoboDojo tests, the authors report that aligned human demonstrations improved out-of-distribution generalization and reduced the target-task robot data needed. Duration and task diversity contributed differently, while video-only supervision remained useful without action labels. The study argues against treating hours alone as the scaling variable; its conclusions remain bounded by the tested data and policy pipeline.
Why it matters
Ego4WAM separated duration, diversity, action labels and human-robot alignment under one world-action model.
Limits and context
- Duration and task diversity contributed differently, while video-only supervision remained useful without action labels.
- The study argues against treating hours alone as the scaling variable; its conclusions remain bounded by the tested data and policy pipeline.
Key claims
Ego4WAM separated duration, diversity, action labels and human-robot alignment under one world-action model.
Qualification: Duration and task diversity contributed differently, while video-only supervision remained useful without action labels.
Evidence: source-2026-10-01-003
Sources
- arXiv preprint 2609.40341arXiv · primary research
Corrections
No corrections have been recorded for this story.