Reproduce “AutoPhyX: Automatic Text-Condition Physics Property Generation” 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

Bei Huang, Yixin Chen, Hongbin Zha, Yuru Pei, Siyuan Huang

AutoPhyX predicts fields of material properties for 3D assets from visual features and a description, using a text embedding to condition a 3D network. It estimates stiffness, Poisson’s ratio, and density in one forward pass. On 30 human-annotated objects, adding text lowers errors across all four measures; this supports text’s value on that test, not recovery of real-world materials.

Summary by alphaXiv.

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