Scalar machine learning of tensorial quantities – Born effective charges from monopole models

Bernhard Schmiedmayer, Angela Rittsteuer, Tobias Hilpert, Georg Kresse · 2026

Read Scalar machine learning of tensorial quantities – Born effective charges from monopole models by Bernhard Schmiedmayer, Angela Rittsteuer, Tobias Hilpert, Georg Kresse (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2602-04773.txt.

Description

Bernhard Schmiedmayer and 3 co-authors predict a tensor property of materials using only scalar machine learning models in a 2026 paper. The target is the Born effective charge tensor, which describes how polarization shifts when atoms move. Tensor targets often demand specially designed tensorial inputs.

Their trick uses the definition: the tensor is a derivative of polarization with respect to atomic displacement. Scalar kernel models, called monopole models here, can thus capture its tensorial nature. The authors compare this with past tensorial dipole kernel models that encode structure directly. Both support charge partitioning into monopole and dipole parts. As a test, they computed finite-temperature infrared spectra for several complex materials.

Catalog record

CREATORBernhard Schmiedmayer, Angela Rittsteuer, Tobias Hilpert, Georg Kresse
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE50 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2602-04773

Browse the full library · About the Tower of Babel