The Many-Body Expansion Combined with Neural Networks

Abstract

Fragmentation methods such as the many-body expansion (MBE) are a common strategy to model large systems by partitioning energies into a hierarchy of decreasingly significant contributions. The number of fragments required for chemical accuracy is still prohibitively expensive for ab-initio MBE to compete with force field approximations for applications beyond single-point energies. Alongside the MBE, empirical models of ab-initio potential energy surfaces have improved, especially non-linear models based on neural networks (NN) which can reproduce ab-initio potential energy surfaces rapidly and accurately. Although they are fast, NNs suffer from their own curse of dimensionality; they must be trained on a representative sample of chemical space. In this paper we examine the synergy of the MBE and NN's, and explore their complementarity. The MBE offers a systematic way to treat systems of arbitrary size and intelligently sample chemical space. NN's reduce, by a factor in excess of $10^6$ the computational overhead of the MBE and reproduce the accuracy of ab-initio calculations without specialized force fields. We show they are remarkably general, providing comparable accuracy with drastically different chemical embeddings. To assess this we test a new chemical embedding which can be inverted to predict molecules with desired properties.

References in corpus (5)

• ANI-1: An extensible neural network potential with DFT accuracy at force field computational cost
• Quantum-Chemical Insights from Deep Tensor Neural Networks
• Machine Learning of Accurate Energy-Conserving Molecular Force Fields
• Comparing the Accuracy of High-Dimensional Neural Network Potentials and the Systematic Molecular Fragmentation Method: A Benchmark Study for all-trans Alkanes
• Computationally Efficient Characterization of Potential Energy Surfaces Based on Fingerprint Distances

Cited by in corpus (24)

• Towards Exact Molecular Dynamics Simulations with Machine-Learned Force Fields
• Machine Learning Molecular Dynamics for the Simulation of Infrared Spectra
• Resolving transition metal chemical space: feature selection for machine learning and structure-property relationships
• Hierarchical modeling of molecular energies using a deep neural network
• WACSF - Weighted Atom-Centered Symmetry Functions as Descriptors in Machine Learning Potentials
• Unified Representation of Molecules and Crystals for Machine Learning
• sGDML: Constructing Accurate and Data Efficient Molecular Force Fields Using Machine Learning
• Predicting Electronic Structure Properties of Transition Metal Complexes with Neural Networks
• Atom-Density Representations for Machine Learning
• Comparison of permutationally invariant polynomials, neural networks, and Gaussian approximation potentials in representing water interactions through many-body expansions
• Molecular Force Fields with Gradient-Domain Machine Learning: Construction and Application to Dynamics of Small Molecules with Coupled Cluster Forces
• Virtual orbital many-body expansions: A possible route towards the full configuration interaction limit
• Metadynamics for Training Neural Network Model Chemistries: a Competitive Assessment
• The Shape of Full Configuration Interaction to Come
• Many-Body Expanded Full Configuration Interaction. I. Weakly Correlated Regime
• Molecular Force Fields with Gradient-Domain Machine Learning (GDML): Comparison and Synergies with Classical Force Fields
• Challenges for Machine Learning Force Fields in Reproducing Potential Energy Surfaces of Flexible Molecules
• Gaussian process model of 51-dimensional potential energy surface for protonated imidazole dimer
• Bypassing the computational bottleneck of quantum-embedding theories for strong electron correlations with machine learning
• Construction of Machine Learned Force Fields with Quantum Chemical Accuracy: Applications and Chemical Insights
• Building nonparametric $n$-body force fields using Gaussian process regression
• Accurate Molecular Dynamics Enabled by Efficient Physically-Constrained Machine Learning Approaches
• Compressing physical properties of atomic species for improving predictive chemistry
• Using a monomer potential energy surface to perform approximate path integral molecular dynamics simulation of ab-initio water with near-zero added cost