Reproduce “AdaFold: Adapting Folding Trajectories of Cloths via Feedback-loop 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

Alberta Longhini, Michael C. Welle, Zackory Erickson, Danica Kragic

AdaFold enables feedback control for quasi-static cloth half-folding by labeling point-cloud layers and repeatedly replanning with a learned dynamics model. It adapts predictions from recent observations and executes one action before replanning. Across 10 simulated cloth-parameter settings, mean final IoU was 0.78 ± 0.11, versus 0.48 ± 0.16 for open-loop AdaFold-OL, while real-world evidence is limited to tested cloths.

Summary by alphaXiv.

Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.