Diffusion probabilistic models enhance variational autoencoder for crystal structure generative modeling.

• DOI: 10.48550/arxiv.2308.02165
• Publication Date: 2024-01-13
• Authors: Teerachote Pakorn...; Natthaphon Choomp...; Sorrjit Arrerut; Chayanon Atthapak; Sakarn Khamkaeo; Thiparat Chotibut; Thiti Bovornratan...
• Keywords: FOS: Computer and information sciences; Computer Science - Machine Learning; Condensed Matter - Materials Science; Statistical Mechanics (cond-mat.stat-mech); Science; Q; R; Materials Science (cond-mat.mtrl-sci); FOS: Physical sciences; Computational Physics (physics.comp-ph); 01 natural sciences; Article; Machine Learning (cs.LG); 0104 chemical sciences; 68T07; Medicine; Physics - Computational Physics; Condensed Matter - Statistical Mechanics

AbstractThe crystal diffusion variational autoencoder (CDVAE) is a machine learning model that leverages score matching to generate realistic crystal structures that preserve crystal symmetry. In this study, we leverage novel diffusion probabilistic (DP) models to denoise atomic coordinates rather than adopting the standard score matching approach in CDVAE. Our proposed DP-CDVAE model can reconstruct and generate crystal structures whose qualities are statistically comparable to those of the original CDVAE. Furthermore, notably, when comparing the carbon structures generated by the DP-CDVAE model with relaxed structures obtained from density functional theory calculations, we find that the DP-CDVAE generated structures are remarkably closer to their respective ground states. The energy differences between these structures and the true ground states are, on average, 68.1 meV/atom lower than those generated by the original CDVAE. This significant improvement in the energy accuracy highlights the effectiveness of the DP-CDVAE model in generating crystal structures that better represent their ground-state configurations.