English

PhysClaw-0: A Symbiotic Agentic System for Robot Autonomy via Language Corrections

Robotics 2026-07-15 v1 Human-Computer Interaction Systems and Control

Abstract

Autonomous data collection governs the volume and quality of real-world trajectories for manipulation policy learning. Existing pipelines reduce human effort via self-resetting, VLM verification, or language-guided correction, yet episode-scoped fixes must be reissued whenever the same failure recurs, so oversight cost grows with session length rather than with the number of distinct problems. We present PhysClaw-0, a human-robot symbiotic agentic system in which corrections are retained and reused across rounds. The collection loop collects, verifies, and resets autonomously, pausing for a remote operator only when a phase exhausts an explicit retry budget. An LLM parser maps each natural-language utterance to a structured adjustment stored in Corrective Memory, so addressed failure modes typically need not be corrected again under the same conditions. On a real-robot desktop-clearing testbed, PhysClaw-0 matches teleoperation episode success while reducing human working time to 16%. Language corrections improve verifier-human agreement in all four evaluated settings and raise average single-attempt success from 12.5% to 47.5% (arm-selection: 20.0% to 50.0%). Policies fine-tuned on PhysClaw-0 data match teleoperation-trained policy success at a fraction of collection human cost.

Cite

@article{arxiv.2607.14047,
  title  = {PhysClaw-0: A Symbiotic Agentic System for Robot Autonomy via Language Corrections},
  author = {Boyuan Wang and Zhenyuan Zhang and Zhiqin Yang and Peijun Gu and Shuya Wang and Xiaofeng Wang and Xianghui Ze and Yifan Chang and Guosheng Zhao and Jiangnan Shao and Guan Huang and Hengyu Liu and Yonggang Zhang and Wei Xue and Chunyuan Guan and Chenglin Pu and Yike Guo and Xingang Wang and Zheng Zhu},
  journal= {arXiv preprint arXiv:2607.14047},
  year   = {2026}
}

Comments

WebPage: https://open-gigaai.github.io/PhysClaw