Representations of molecules and materials for interpolation of quantum-mechanical simulations via machine learning

Abstract

Computational study of molecules and materials from first principles is a cornerstone of physics, chemistry, and materials science, but limited by the cost of accurate and precise simulations. In settings involving many simulations, machine learning can reduce these costs, often by orders of magnitude, by interpolating between reference simulations. This requires representations that describe any molecule or material and support interpolation. We comprehensively review and discuss current representations and relations between them, using a unified mathematical framework based on many-body functions, group averaging, and tensor products. For selected state-of-the-art representations, we compare energy predictions for organic molecules, binary alloys, and Al-Ga-In sesquioxides in numerical experiments controlled for data distribution, regression method, and hyper-parameter optimization.

20 pages, 6 figures, excluding supplement (19 pages, 5 figures); v2: extended review and discussion, more representations covered, edited for clarity. For additional information, including datasets, results, and software see https://marcel.science/repbench

References in corpus (38)

• ANI-1: An extensible neural network potential with DFT accuracy at force field computational cost
• Kernel methods in machine learning
• A Performance and Cost Assessment of Machine Learning Interatomic Potentials
• Machine Learning Unifies the Modelling of Materials and Molecules
• On-the-fly machine learning force field generation: Application to melting points
• Phase transitions of hybrid perovskites simulated by machine-learning force fields trained on-the-fly with Bayesian inference
• A Fourth-Generation High-Dimensional Neural Network Potential with Accurate Electrostatics Including Non-local Charge Transfer
• Physics-inspired structural representations for molecules and materials
• Deep neural network solution of the electronic Schrödinger equation
• SchNet: A continuous-filter convolutional neural network for modeling quantum interactions
• FCHL revisited: faster and more accurate quantum machine learning
• Symmetry-Adapted Machine-Learning for Tensorial Properties of Atomistic Systems
• Understanding molecular representations in machine learning: The role of uniqueness and target similarity
• OrbNet: Deep Learning for Quantum Chemistry Using Symmetry-Adapted Atomic-Orbital Features
• Incorporating long-range physics in atomic-scale machine learning
• Combining SchNet and SHARC: The SchNarc machine learning approach for excited-state dynamics
• On the Completeness of Atomic Structure Representations
• Atomic cluster expansion of scalar, vectorial and tensorial properties and including magnetism and charge transfer
• A Universal Density Matrix Functional from Molecular Orbital-Based Machine Learning: Transferability across Organic Molecules
• Gaussian Moments as Physically Inspired Molecular Descriptors for Accurate and Scalable Machine Learning Potentials
• Ground state energy functional with Hartree-Fock efficiency and chemical accuracy
• Recursive evaluation and iterative contraction of $N$-body equivariant features
• Deep Learning for UV Absorption Spectra with SchNarc: First Steps Towards Transferability in Chemical Compound Space
• Neural networks and kernel ridge regression for excited states dynamics of CH${}_2$NH${}_2^{+}$: From single-state to multi-state representations and multi-property machine learning models
• Machine learning and excited-state molecular dynamics
• Permutationally Invariant, Reproducing Kernel-Based Potential Energy Surfaces for Polyatomic Molecules: From Formaldehyde to Acetone
• Improved accuracy and transferability of molecular-orbital-based machine learning: Organics, transition-metal complexes, non-covalent interactions, and transition states
• Group-theoretical high-order rotational invariants for structural representations: Application to linearized machine learning interatomic potential
• Comment on "Fast and Accurate Modeling of Molecular Atomization Energies with Machine Learning"
• A Novel Approach to Describe Chemical Environments in High Dimensional Neural Network Potentials
• An assessment of the structural resolution of various fingerprints commonly used in machine learning
• Efficient implementation of atom-density representations
• Machine learning potentials for multicomponent systems: The Ti-Al binary system
• ML Models of Vibrating H${}_2$CO: Comparing Reproducing Kernels, FCHL and PhysNet
• Sensitivity and Dimensionality of Atomic Environment Representations used for Machine Learning Interatomic Potentials
• Through the eyes of a descriptor: Constructing complete, invertible descriptions of atomic environments
• Wavelet Scattering Networks for Atomistic Systems with Extrapolation of Material Properties
• Analytical Gradients for Molecular-Orbital-Based Machine Learning

Cited by in corpus (30)

• GPUMD: A package for constructing accurate machine-learned potentials and performing highly efficient atomistic simulations
• Neural Network Potentials for Chemistry: Concepts, Applications and Prospects
• Updates to the DScribe Library: New Descriptors and Derivatives
• High-Dimensional Neural Network Potentials for Magnetic Systems Using Spin-Dependent Atom-Centered Symmetry Functions
• Ultra-fast interpretable machine-learning potentials
• Uncertainty-biased molecular dynamics for learning uniformly accurate interatomic potentials
• Predicting tensorial molecular properties with equivariant machine-learning models
• Toward Accurate Interpretable Predictions of Materials Properties within Transformer Language Models
• The MD17 Datasets from the Perspective of Datasets for Gas-Phase "Small" Molecule Potentials
• Advances in modeling complex materials: The rise of neuroevolution potentials
• KLIFF: A framework to develop physics-based and machine learning interatomic potentials
• 3DReact: Geometric deep learning for chemical reactions
• Kernel based quantum machine learning at record rate : Many-body distribution functionals as compact representations
• Stress and heat flux via automatic differentiation
• Ab initio machine learning of phase space averages
• Machine Learning for compositional disorder: A Comparison Between Different Descriptors and Machine Learning Frameworks
• Towards Structural Reconstruction from X-Ray Spectra
• Mean-Field Density Matrix Decompositions
• Probing the effects of broken symmetries in machine learning
• SPA${}^{\mathrm{H}}$M(a,b): encoding the density information from guess Hamiltonian in quantum machine learning representations
• Atomistic Simulations of Oxide-Water Interfaces using Machine Learning Potentials
• Structural Descriptors and Information Extraction from X-ray Emission Spectra: Aqueous Sulfuric Acid
• Ab initio machine learning in chemical compound space
• MADAS -- A Python framework for assessing similarity in materials-science data
• Preparing Quantum States by Measurement-feedback Control with Bayesian Optimization
• Revving up 13C NMR shielding predictions across chemical space: Benchmarks for atoms-in-molecules kernel machine learning with new data for 134 kilo molecules
• Self-supervised Representations and Node Embedding Graph Neural Networks for Accurate and Multi-scale Analysis of Materials
• Efficient interpolation of molecular properties across chemical compound space with low-dimensional descriptors
• Unified theory of atom-centered representations and message-passing machine-learning schemes
• Integer linear programming for unsupervised training set selection in molecular machine learning