Reproduce “PhysEvo: Astra Can Act, Let It” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Wenqing Tian, Zeyu Zhang, Zhaocheng Liu, Fengwei Liu, Qiang Liu, Liang Wang
PhysEvo improves robot control by letting a frozen Astra act while a meta-agent reviews execution evidence and revises task-specific tools and skills. It tests and retains useful changes between episodes. Across 42 RoboDojo tasks, retained task-specific versions achieve 68.14/100 and 62.00% success, versus RoboDawn’s published 47.17% one-shot success rate. These results concern the reported simulation protocol, not a paired comparison.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.