Reproduce “DexHOT: Learning Hand-Object Configuration Transfer for Dexterous Manipulation” 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
Zhewei Gui, Anthony Bolton, Jiyao Zhang, Zeyuan Chen, Qiyang Yan, Tianyu Wang, Hao Dong
DexHOT learns reusable hand–object configuration transfers, specifying hand articulation and object pose relative to the wrist. A shared policy is trained on feasible configurations and follows planner- or demonstration-guided subgoals. In DexHOTBench it achieves 80.5% success on seen objects versus 69.4% among 80 held-out objects, with lower scores for unseen-category objects and far transfers.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.