Reproduce “Human-to-Robot Transfer via Bridging Representation and Action Space in VLAs” 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
Junhyeong Park, Wook Jung, Taeyoung Kim, Huiwon Jang, Jinwoo Shin
H2RB transfers target robot tasks from human demonstrations without paired human–robot examples. It grounds visual features in shared hand geometry and predicts shared latent actions decoded into human- and robot-specific motion. Source robot experience remains in training. On OpenArm, it scores 62.7 average task progress on Open and Close, versus 0.0 for a matched baseline; object transfer scores are 51.8 versus 8.0. These are progress scores, not success rates.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.