Reproduce “ReDex: Repairing Sim-to-Real Dexterous Policies by Finger-Level Compliant Interaction” 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
Jinzhou Li, Hadi Tabatabaee, Kelin Yu, Yuyin Sun, Cheng-Hao Kuo, Roberto Martín-Martín, Nima Fazeli, X. Alice Wu and 1 more
ReDex adapts a simulated dexterous-hand policy through physical corrections to failing fingers while the frozen policy controls others. It records fingertip forces and converts them into force-informed targets for a standalone policy conditioned on joint motion and force. On XHAND1 it achieves 86% flipping success versus 14% for RMA and 95.3% screwdriver progress versus 26.0%; findings cover two skills on one platform.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.