Reproduce “GRASP: Gradient-Guided Sampling for Contact-Rich Physics-Based Motion Retargeting” 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
Yang Liu, Tong Wang, Yiming Bao, Heng Yang
GRASP guides trajectory search using previously evaluated simulator costs rather than derivatives through contact dynamics. For long motions, it replans after executing part of a plan. In bimanual retargeting of 1,352 OakInk2 demonstrations to Sharpa hands, it achieved 92.8% success at 1,024 particles, versus 81.0% for Do-as-I-Do, under the paper’s object-tracking thresholds; this result is specific to that evaluation.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.