An equivariant message-passing neural network embeds atomic spins explicitly to learn magnetic interactions, achieving near-DFT accuracy and data efficiency across magnetic systems via fine-tuning.
We first per- formed a reference density-functional theory (DFT) cal- culation by fully relaxing the structure initialized accord- ing to Ref
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cond-mat.mtrl-sci 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Equivariant Many-body Message Passing Interatomic Potentials for Magnetic Materials
An equivariant message-passing neural network embeds atomic spins explicitly to learn magnetic interactions, achieving near-DFT accuracy and data efficiency across magnetic systems via fine-tuning.