Exploring the design space of machine-learning models for quantum chemistry with a fully differentiable framework

Abstract

Traditional atomistic machine learning (ML) models serve as surrogates for quantum mechanical (QM) properties, predicting quantities such as dipole moments and polarizabilities, directly from compositions and geometries of atomic configurations. With the emergence of ML approaches to predict the "ingredients" of a QM calculation, such as the ground state charge density or the effective single-particle Hamiltonian, it has become possible to obtain multiple properties through analytical physics-based operations on these intermediate ML predictions. We present a framework to seamlessly integrate the prediction of an effective electronic Hamiltonian, for both molecular and condensed-phase systems, with PySCFAD, a differentiable QM workflow that facilitates its indirect training against functions of the Hamiltonian, such as electronic energy levels, dipole moments, polarizability, etc. We then use this framework to explore various possible choices within the design space of hybrid ML/QM models, examining the influence of incorporating multiple targets on model performance and learning a reduced-basis ML Hamiltonian that can reproduce targets computed from a much larger basis. Our benchmarks evaluate the accuracy and transferability of these hybrid models, compare them against predictions of atomic properties from their surrogate models, and provide indications to guide the design of the interface between the ML and QM components of the model.

References in corpus (36)

• Phonons and related properties of extended systems from density-functional perturbation theory
• Gaussian Approximation Potentials: the accuracy of quantum mechanics, without the electrons
• On representing chemical environments
• Fast and Accurate Modeling of Molecular Atomization Energies with Machine Learning
• E(3)-Equivariant Graph Neural Networks for Data-Efficient and Accurate Interatomic Potentials
• A Spectral Analysis Method for Automated Generation of Quantum-Accurate Interatomic Potentials
• Machine Learning of Molecular Electronic Properties in Chemical Compound Space
• Ab-Initio Solution of the Many-Electron Schrödinger Equation with Deep Neural Networks
• Symmetry-Adapted Machine-Learning for Tensorial Properties of Atomistic Systems
• Automatic Selection of Atomic Fingerprints and Reference Configurations for Machine-Learning Potentials
• A Transferable Machine-Learning Model of the Electron Density
• OrbNet: Deep Learning for Quantum Chemistry Using Symmetry-Adapted Atomic-Orbital Features
• Deep-Learning Density Functional Theory Hamiltonian for Efficient ab initio Electronic-Structure Calculation
• Transferability in Machine Learning for Electronic Structure via the Molecular Orbital Basis
• Accurate molecular polarizabilities with coupled-cluster theory and machine learning
• Deep neural network for the dielectric response of insulators
• Ab-initio quantum chemistry with neural-network wavefunctions
• General framework for E(3)-equivariant neural network representation of density functional theory Hamiltonian
• Predicting molecular dipole moments by combining atomic partial charges and atomic dipoles
• Automatic differentiation in quantum chemistry with an application to fully variational Hartree-Fock
• Transferable E(3) equivariant parameterization for Hamiltonian of molecules and solids
• Recursive evaluation and iterative contraction of $N$-body equivariant features
• Universal Machine Learning for the Response of Atomistic Systems to External Fields
• Physically inspired deep learning of molecular excitations and photoemission spectra
• Equivariant representations for molecular Hamiltonians and N-center atomic-scale properties
• Equivariant analytical mapping of first principles Hamiltonians to accurate and transferable materials models
• Message-Passing Neural Quantum States for the Homogeneous Electron Gas
• Differentiable quantum chemistry with PySCF for molecules and materials at the mean-field level and beyond
• Universal Machine Learning Kohn-Sham Hamiltonian for Materials
• DeepQMC: an open-source software suite for variational optimization of deep-learning molecular wave functions
• SPA${}^{\mathrm{H}}$M: the Spectrum of Approximated Hamiltonian Matrices representations
• Automatic Differentiation for Orbital-Free Density Functional Theory
• Exploring the design space of machine-learning models for quantum chemistry with a fully differentiable framework
• Learning on-top: regressing the on-top pair density for real-space visualization of electron correlation
• Molecular Dipole Moment Learning via Rotationally Equivariant Gaussian Process Regression with Derivatives in Molecular-orbital-based Machine Learning
• Integrated workflows and interfaces for data-driven semi-empirical electronic structure calculations

Cited by in corpus (6)

• Exploring the design space of machine-learning models for quantum chemistry with a fully differentiable framework
• Metatensor and metatomic: foundational libraries for interoperable atomistic machine learning
• Weighted Active Space Protocol for Multireference Machine-Learned Potentials
• Comparing the latent features of universal machine-learning interatomic potentials
• Machine learning of electronic structure and atomistic properties from the external potential
• Predictive Free Energy Simulations Through Hierarchical Distillation of Quantum Hamiltonians