Reproduce “REVOLVE: An Automated Closed-Loop Framework for Evolving Robot Manipulation with Minimal Human Intervention” 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
Hanyu Liu, Qian Li, Yizhu Ding, Jiayi Wen, Keqiang Ren, Yunsheng Ma, Tao Jian, Zhihua Wang and 3 more
REVOLVE automates robot reset and recovery, then feeds successful runs and corrected failures into policy training while storing human-verified supervisory disagreements for retrieval-based judgment. Across four tabletop tasks, it reduced human time for 400 collection trials from 8.89 to 0.50 hours and 400 deployment tests from 2.67 to 0.13 hours. After five updates, success averaged 58.5% versus 45%; complex-task data acquisition remained human-led.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.