Reproduce “Learning from Fleet Experience: Collective Experience Learning Across 96 Robots” 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
Chao Yu, Shu'ang Yu, Hao Lin, Hongzhi Zang, Yuxiang Yan, Chenxi Jiang, Zhen Guo, Chuankang Li and 36 more
FERT lets robots improve task-specific policies from fleet experience after expert-supervised initialization. It uses failed episodes to retrieve similar moments from successful runs, then emphasizes those recorded actions in supervised training; operators still label outcomes and reset environments. On phone packaging, success rose from 79.2% to 98.3% with 96 robots, tested on held-out AgiBot G2 instances.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.