Reproduce “In-Context Learning World Models for Robot Dual-Arm Manipulation of Soft Garment with Partial Observations” 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

Hengyi Sim, Quentin Rouxel, Dennis Hong, Darwin Caldwell, Fei Chen

The authors present a system for garment folding from partly obscured views when fabric response is uncertain. It reconstructs point clouds from partial depth observations, uses interactions as context to predict candidate folds, and selects actions by predicted outcomes without updating model weights. In real-world trials, 55 of 60 rollouts met a success threshold; evidence is limited to primarily quasi-static folding.

Summary by alphaXiv.

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