Machine Learning Magnetism Classifiers from Atomic Coordinates

Published in iScience, 2022

Paper /

A Euclidean-equivariant graph neural network classifies magnetic order directly from atomic coordinates, using a descriptor that encodes crystal structures of any space group, as a data-driven complement to neutron diffraction and semi-empirical DFT.

Recommended citation: Merker, H.A., Heiberger, H., Nguyen, L., Liu, T., Chen, Z., Andrejevic, N., Drucker, N.C., Okabe, R., Wang, Y., Smidt, T., & Li, M. (2022). Machine Learning Magnetism Classifiers from Atomic Coordinates. iScience 25, 10, 105192.
Download Paper