Reproduce “Never Too Late for Force: Accelerating VLA Post-Training with Reactive Force Injection” 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
Yi Wang, Wendi Chen, Zimo Wen, Han Xue, Xueqi Li, Wenye Yu, Zhijie Chen, Hao Yang and 3 more
LIFT adapts a pretrained vision-led robot policy for contact, adding a force-responsive action path and training on corrections from states the robot visits. It encodes recent force history to refresh actions during contact. On book insertion, LIFT reaches a peak mean score of 0.6, versus 0.4 for its single-frame-force variant; towel folding is an exception, where performance is comparable.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.