NEEDLEWORK: Offline Rewriting of Robot Data with Verified Local Stitches
NEEDLEWORK: Offline Rewriting of Robot Data with Verified Local Stitches
This paper presents NEEDLEWORK, an offline dataset-augmentation algorithm that improves robot training data by adding verified action bridges between recorded observations in high-dimensional demonstrations. The method uses RGB images, proprioception, and episode-level outcomes to create connections that bypass suboptimal detours, broaden action coverage, and include failed trajectories. It enhances success rates on real-robot tasks by an average of 21 percentage points over strong baselines.
Full article
You are reading the complete RoboSignal summary. The publisher’s full article is available at the original source.
Read full article at sourcearxiv.org · Opens in a new tab; source language may differ.
What the source reports
Publisher-reported claims, with original evidence. These results have not been independently verified by RoboSignal.
What remains unknown
Not established in the collected evidence: Environment, Control, Data origin.
Reported performance applies to the described task. It does not establish general autonomy or deployment readiness.
Source excerpts and review record
Automatically extracted; no manual editorial approval recorded.
Implications for data suppliers
RoboSignal interpretation and collection questions, not statements of buyer demand.
- Confirm the required data type and collection setting with the buyer; this source does not establish a complete collection specification.
- Validate demand and acceptance criteria with a buyer before scaling. Publication, popularity and a research result do not establish a purchase commitment.
Source：arXiv Robotics — research abstracts · arxiv.org