arXivChenzhi Liu, Yue Zhang, Jiehong Lin, Jianan Wang, Bo Wang, Zhongrui Wang, Xiaojuan QiFri, Oct 2, 2026, 8:48 AM PDT
score 15.1
Robot agent learns from its own mistakes to handle long tasks
Original: MobiAgent: Dual-Loop Recursive Policy Self-Improvement for Long-Horizon Mobile Manipulation
Source: arxiv.org ↗
Writing ELI5 summary…