MEKAWIRE
ExpertarXiv

SEAR: Self-Evolving Agentic Robotics

Xinyi Wang, Bangzheng Li, Hanyu Wang, Lin Zhang, Andrea Yaoyun Cui, Zhenggang Tang, Shyamal Buch, Dejia Xu
Original paper Read the full paper (PDF)
From the abstract

Frontier language-model agents can now improve robots on their own: they write control code, test it, and keep what works as skills and memories for the next task. We introduce SEAR (Self-Evolving Agentic Robotics), a framework and benchmark built around two questions.

From the abstract. Our summary is in progress.

The full text is not republished here because the paper's license does not allow it. Read the original on arXiv.