Reproduce “CSO: Refining Robotic Policies via Skill Distribution Alignment and Skill-Grained Optimization” 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
Zhiyuan Xiang, Xiang Deng, Wanjie Tao, Xu Chen, Zhao Jielun, Weili Guan
CSO improves robotic policies by learning from successful executions and optimizing selection of complete motion skills. It keeps the action tokenizer and decoder fixed, while refining the observation-to-skill policy. On LIBERO, its overall success rate rises from 80.1% to 97.5%. This is an unweighted mean across four suites; performance remains bounded by available initial successes and tokenizer motion coverage.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.