Reproduce “REBOOT: From Failure to Recovery - A Dataset and Benchmark for Precision Assembly” 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
Nana Yaw Owusu Ofori-Ampofo, Samira Ebrahimi Kahou, Joseph Thekinen
REBOOT presents a real-robot precision-assembly dataset and benchmark with 2,160 demonstrations across 18 tasks, including expert successes and teleoperated corrections of policy failures. It annotates five task stages, failure modes, mechanical metadata, language, and synchronized RGB-D views. Across 810 analyzed rollouts, placement stages account for 54% of installation failures, while installation success averages 16% in this controlled suite.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.