arXiv: Asynchronous Robot Policies
Scaling vision language foundation models has significantly advanced generalist robot manipulation, but the increased model scale introduces computational latency that degrades real time physical reactivity.
The arXiv Coding-Agent Research page documents the arXiv update.