A Recipe for Charge Density Prediction

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Fu, Xiang, Rosen, Andrew, Bystrom, Kyle, Wang, Rui, Musaelian, Albert, Kozinsky, Boris, Smidt, Tess, Jaakkola, Tommi
Natura: Preprint
Pubblicazione: 2024
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866917678406434816
author Fu, Xiang
Rosen, Andrew
Bystrom, Kyle
Wang, Rui
Musaelian, Albert
Kozinsky, Boris
Smidt, Tess
Jaakkola, Tommi
author_facet Fu, Xiang
Rosen, Andrew
Bystrom, Kyle
Wang, Rui
Musaelian, Albert
Kozinsky, Boris
Smidt, Tess
Jaakkola, Tommi
contents In density functional theory, charge density is the core attribute of atomic systems from which all chemical properties can be derived. Machine learning methods are promising in significantly accelerating charge density prediction, yet existing approaches either lack accuracy or scalability. We propose a recipe that can achieve both. In particular, we identify three key ingredients: (1) representing the charge density with atomic and virtual orbitals (spherical fields centered at atom/virtual coordinates); (2) using expressive and learnable orbital basis sets (basis function for the spherical fields); and (3) using high-capacity equivariant neural network architecture. Our method achieves state-of-the-art accuracy while being more than an order of magnitude faster than existing methods. Furthermore, our method enables flexible efficiency-accuracy trade-offs by adjusting the model/basis sizes.
format Preprint
id arxiv_https___arxiv_org_abs_2405_19276
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle A Recipe for Charge Density Prediction
Fu, Xiang
Rosen, Andrew
Bystrom, Kyle
Wang, Rui
Musaelian, Albert
Kozinsky, Boris
Smidt, Tess
Jaakkola, Tommi
Computational Physics
Machine Learning
In density functional theory, charge density is the core attribute of atomic systems from which all chemical properties can be derived. Machine learning methods are promising in significantly accelerating charge density prediction, yet existing approaches either lack accuracy or scalability. We propose a recipe that can achieve both. In particular, we identify three key ingredients: (1) representing the charge density with atomic and virtual orbitals (spherical fields centered at atom/virtual coordinates); (2) using expressive and learnable orbital basis sets (basis function for the spherical fields); and (3) using high-capacity equivariant neural network architecture. Our method achieves state-of-the-art accuracy while being more than an order of magnitude faster than existing methods. Furthermore, our method enables flexible efficiency-accuracy trade-offs by adjusting the model/basis sizes.
title A Recipe for Charge Density Prediction
topic Computational Physics
Machine Learning
url https://arxiv.org/abs/2405.19276