A general formalism for machine-learning models based on multipolar-spherical harmonics

Abstract

The formulation of descriptors of the local chemical environment, enabling the construction of machine-learning models, is usually obtained by studying the properties of the expansion coefficients of a neighborhood density. In this work, we show that all the transformation properties of the descriptors and their behaviour under rotation, inversion and complex conjugation, are derived from the choice of the basis over which the density is expanded. Furthermore, crucially they are independent from the explicit mathematical form of the neighborhood density. In particular, we show that all the descriptors investigated, can be obtained by an expansion in multipolar spherical harmonics, which constitutes the core of this work, and which is introduced and analysed in great detail. By exploiting the orthogonality and the transformation rules of the multipolar spherical harmonics, we show that several formulations are simplified, such as the one needed to obtain the $\lambda -$SOAP kernel and its properties. We close this work by applying our framework to several multi-body descriptors available in literature, providing an in-depth analysis of their main properties, as made clear from the vantage viewpoint of a basis-centered approach.

28 pages, 2 figures

References in corpus (30)

• Gaussian Approximation Potentials: the accuracy of quantum mechanics, without the electrons
• On representing chemical environments
• DeePMD-kit: A deep learning package for many-body potential energy representation and molecular dynamics
• ANI-1: An extensible neural network potential with DFT accuracy at force field computational cost
• Moment Tensor Potentials: a class of systematically improvable interatomic potentials
• A Spectral Analysis Method for Automated Generation of Quantum-Accurate Interatomic Potentials
• A Universal Graph Deep Learning Interatomic Potential for the Periodic Table
• A Performance and Cost Assessment of Machine Learning Interatomic Potentials
• Comparing molecules and solids across structural and alchemical space
• Machine-learning based interatomic potential for amorphous carbon
• Physics-inspired structural representations for molecules and materials
• Machine learning a general purpose interatomic potential for silicon
• Symmetry-Adapted Machine-Learning for Tensorial Properties of Atomistic Systems
• Chemical Shifts in Molecular Solids by Machine Learning
• An Accurate and Transferable Machine Learning Potential for Carbon
• Accurate Interatomic Force Fields via Machine Learning with Covariant Kernels
• Measuring Statistical isotropy of the CMB anisotropy
• Atom-Density Representations for Machine Learning
• Atomic cluster expansion of scalar, vectorial and tensorial properties and including magnetism and charge transfer
• Recursive evaluation and iterative contraction of $N$-body equivariant features
• Explicit Multi-element Extension of the Spectral Neighbor Analysis Potential for Chemically Complex Systems
• Accelerating Finite-temperature Kohn-Sham Density Functional Theory with Deep Neural Networks
• Equivariant analytical mapping of first principles Hamiltonians to accurate and transferable materials models
• Predicting tensorial molecular properties with equivariant machine-learning models
• A spectral-neighbour representation for vector fields: machine-learning potentials including spin
• Isotropic N-Point Basis Functions and Their Properties
• Bipolar Harmonic encoding of CMB correlation patterns
• Linear Jacobi-Legendre expansion of the charge density for machine learning-accelerated electronic structure calculations
• Multipole expansions in four-dimensional hyperspherical harmonics
• High-performance descriptor for magnetic materials: Accurate discrimination of magnetic structure

Cited by in corpus (1)

• Representing spherical tensors with scalar-based machine-learning models