Magnetic HIP-NN for spin dynamics in disordered itinerant magnets
Researchers have developed a new magnetic extension to the Hierarchically Interacting Particle Neural Network (HIP-NN), termed mHIP-NN. This advancement allows for large-scale simulations of spin dynamics in disordered itinerant magnets by directly incorporating rotationally invariant spin correlations into the network's message-passing layers. The mHIP-NN can accurately learn emergent magnetic energy landscapes and effective local fields, proving effective in simulating complex spin dynamics that are computationally intensive with traditional methods. AI
IMPACT Introduces a novel neural network architecture for simulating complex magnetic phenomena, potentially accelerating materials science research.