Reproduce “Equivariant Neural Operator Learning with Graphon Convolution” 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
Chaoran Cheng, Jian Peng
InfGCN predicts electron density as a continuous three-dimensional field from molecular graphs. It combines atom-centered Gaussian basis functions with geometry-aware message passing and a coordinate-specific residual that adjusts estimates at queried points, while preserving SE(3)-equivariance. On randomly rotated QM9, it reports 4.73% NMAE, versus 4.98% for DeepDFT2. This measures density-prediction error under the tested protocol, not computational speedup.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.