Reproduce “DEXTERA: From a Single Image to Deployable Dexterous Manipulation via Real-to-Sim-to-Real” 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
Jin Wu, Lianjie Yuan, Zeyan Sun, Yuanyuan Lei, Disi A, Bicheng Han, Fangzhou Xia
DEXTERA builds an aligned interactive simulation from a workspace image and connects it to dexterous robot learning. Asset generation uses the image, but metric alignment requires a robot model and calibration data. Hardware replay succeeded for 343/390 adapted simulation-successful trajectories (87.95%). Separately, co-training raised mean policy success from 29.2% to 61.9% across six task–embodiment pairs; this is not zero-shot transfer.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.