Reproduce “TacClaw-Infinity: A Large-Scale Vision-Tactile-Language-Action Dataset 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
Fucai Zhu, Xidong Zhang, Yichi Zhang, Yipan Wang, Fubo Wang, Zibin Wu, Kaiqi Li, Jianhao Jiao and 18 more
TacClaw-Infinity presents a robot-independent dataset of contact-rich demonstrations collected with a handheld tactile gripper. It synchronizes vision, touch, motion, and language, using contact events to structure temporal phases and spatial annotations. Policies tested on three tasks and three robots show π₀.₅’s average success rate rises from 50% to 74% with tactile input, and gains vary by task.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.