research
The Video Model Must Say Which Hand Did What—and When
EgoActionBench turns continuous first-person manipulation into 41,132 human-refined atomic actions with hand identity and temporal boundaries.
Summary
EgoActionBench turns continuous first-person manipulation into 41,132 human-refined atomic actions with hand identity and temporal boundaries.
Broad video captions can miss the fine structure needed to learn manipulation. EgoActionBench labels 1,407 first-person episodes with 41,132 atomic actions assigned to the left hand, right hand or both, then scores semantic fidelity, timing, precision and recall. The authors benchmark 20 systems and adapt two open vision-language backbones with supervised learning followed by reinforcement learning. Their adapted models gain 15.9 and 23.2 points on the proposed EgoActionScore and beat the tested open-source baselines. The benchmark also exposes continuing failures in action atomicity, temporal precision and hand association; its score is a research measure, not evidence that a model understands arbitrary human activity.
Why it matters
EgoActionBench turns continuous first-person manipulation into 41,132 human-refined atomic actions with hand identity and temporal boundaries.
Limits and context
- The benchmark also exposes continuing failures in action atomicity, temporal precision and hand association; its score is a research measure, not evidence that a model understands arbitrary human activity.
Key claims
EgoActionBench turns continuous first-person manipulation into 41,132 human-refined atomic actions with hand identity and temporal boundaries.
Qualification: The benchmark also exposes continuing failures in action atomicity, temporal precision and hand association; its score is a research measure, not evidence that a model understands arbitrary human activity.
Evidence: source-2026-11-03-003
Sources
- arXiv preprint 2611.03646, original v1arXiv · primary research
Corrections
No corrections have been recorded for this story.