Parts | Days | Selection | Search | Updates | Downloads | Help

O: Fachverband Oberflächenphysik

O 43: Frontiers of Electronic Structure Theory: Focus on Artificial Intelligence Applied to Real Materials 1

O 43.10: Talk

Wednesday, September 7, 2022, 12:45–13:00, S054

Equivariant graph neural network for linear scaling electron density estimation and applications in battery materials — •Arghya Bhowmik and Peter Jorgensen — 301 Anker Engelunds vej, Kgs. Lyngby, DK-2800

We present a machine learning framework for the prediction of ρ(r) based on equivariant graph message passing neural networks. The electron density is predicted at special query point vertices that are part of the message passing graph, but only receive messages. The model is tested across multiple data sets of molecules (QM9), liquid ethylene carbonate electrolyte (EC) and LixNiyMnzCo(1-y-z)O2 lithium ion battery cathodes (NMC). The model is used to explore large materials phase space for safer battery materials and uncovering new understanding how redox mediated diffusion occurs and battery materials.