Metadynamics for Training Neural Network Model Chemistries: a Competitive Assessment

Abstract

Neural network (NN) model chemistries (MCs) promise to facilitate the accurate exploration of chemical space and simulation of large reactive systems. One important path to improving these models is to add layers of physical detail, especially long-range forces. At short range, however, these models are data driven and data limited. Little is systematically known about how data should be sampled, and `test data' chosen randomly from some sampling techniques can provide poor information about generality. If the sampling method is narrow `test error' can appear encouragingly tiny while the model fails catastrophically elsewhere. In this manuscript we competitively evaluate two common sampling methods: molecular dynamics (MD), normal-mode sampling (NMS) and one uncommon alternative, Metadynamics (MetaMD), for preparing training geometries. We show that MD is an inefficient sampling method in the sense that additional samples do not improve generality. We also show MetaMD is easily implemented in any NNMC software package with cost that scales linearly with the number of atoms in a sample molecule. MetaMD is a black-box way to ensure samples always reach out to new regions of chemical space, while remaining relevant to chemistry near $k_{b}T$. It is one cheap tool to address the issue of generalization.

References in corpus (13)

• ANI-1: An extensible neural network potential with DFT accuracy at force field computational cost
• Machine-learning based interatomic potential for amorphous carbon
• How van der Waals interactions determine the unique properties of water
• Equilibrium free energies from non-equilibrium metadynamics
• Massively Multitask Networks for Drug Discovery
• ANI-1: A data set of 20M off-equilibrium DFT calculations for organic molecules
• Resolving transition metal chemical space: feature selection for machine learning and structure-property relationships
• Symmetry-Adapted Machine-Learning for Tensorial Properties of Atomistic Systems
• Hierarchical modeling of molecular energies using a deep neural network
• Materials Cartography: Representing and Mining Material Space Using Structural and Electronic Fingerprints
• On the accuracy of the MB-pol many-body potential for water: Interaction energies, vibrational frequencies, and classical thermodynamic and dynamical properties from clusters to liquid water and ice
• Non-covalent interactions across organic and biological subsets of chemical space: Physics-based potentials parametrized from machine learning
• Atomic Convolutional Networks for Predicting Protein-Ligand Binding Affinity

Cited by in corpus (19)

• Machine Learning Force Fields
• Less is more: sampling chemical space with active learning
• Active Learning of Uniformly Accurate Inter-atomic Potentials for Materials Simulation
• Machine learning for electronically excited states of molecules
• Silicon liquid structure and crystal nucleation from ab-initio deep Metadynamics
• Guest Editorial: Special Topic on Data-enabled Theoretical Chemistry
• Machine learning and excited-state molecular dynamics
• Deep Coarse-grained Potentials via Relative Entropy Minimization
• Fast Neural Network Approach for Direct Covariant Forces Prediction in Complex Multi-Element Extended Systems
• Deep Neural Network Computes Electron Densities and Energies of a Large Set of Organic Molecules Faster than Density Functional Theory (DFT)
• Metadynamics sampling in atomic environment space for collecting training data for machine learning potentials
• Towards fully ab initio simulation of atmospheric aerosol nucleation
• Molecular Quantum Chemical Data Sets and Databases for Machine Learning Potentials
• Neural Networks Potential from the Bispectrum Component: A Case Study on Crystalline Silicon
• Physics-informed active learning for accelerating quantum chemical simulations
• Weighted Active Space Protocol for Multireference Machine-Learned Potentials
• Compressing physical properties of atomic species for improving predictive chemistry
• The Homunculus Brain and Categorical Logic
• Incorporating electronic information into Machine Learning potential energy surfaces via approaching the ground-state electronic energy as a function of atom-based electronic populations