Reproduce “ForeTac: Predictive Contact Guidance for Generative Robot Policies” 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
Zhiyuan Chen, Shuai Chen, Haoran Zhou, Yizhe Zhao, Kaiyuan Yang, Zhixiang Duan, Shaoqing Xu, Ran Cao and 2 more
ForeTac steers diffusion and flow-matching policies by predicting tactile consequences of candidate actions and backpropagating contact-quality scores into sampling. Across four equally weighted real-robot tasks, adding ForeTac to Diffusion Policy raised average task success from 36.3% to 71.3% and contact success from 20.0% to 66.3%. The task-specific scorer limits evidence to these tasks.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.