Reproduce “SU(2)-Equivariant Spinor Message Passing for Molecular Magnetism” 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
Erh-Wei Sheng, Beom Seok Kang, Yuchao Lin, Danish Khan, Anima Anandkumar
eSpinor predicts crystal-field coefficients for molecular geometries by propagating half-integer spinor features through message passing and contracting them into physical predictions. For the Dy(III) PARASHIFT probe, its real electron-spin correlation error at 8,000 training geometries is 0.553%, versus 1.140% for NequIP+H, against CASSCF-SO. This advantage is specific to the evaluated real correlation; imaginary correlations and fitted decay parameters do not rank the models likewise.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.