WACSF - Weighted Atom-Centered Symmetry Functions as Descriptors in Machine Learning Potentials

Abstract

We introduce weighted atom-centered symmetry functions (wACSFs) as descriptors of a chemical system's geometry for use in the prediction of chemical properties such as enthalpies or potential energies via machine learning. The wACSFs are based on conventional atom-centered symmetry functions (ACSFs) but overcome the undesirable scaling of the latter with increasing number of different elements in a chemical system. The performance of these two descriptors is compared using them as inputs in high-dimensional neural network potentials (HDNNPs), employing the molecular structures and associated enthalpies of the 133855 molecules containing up to five different elements reported in the QM9 database as reference data. A substantially smaller number of wACSFs than ACSFs is needed to obtain a comparable spatial resolution of the molecular structures. At the same time, this smaller set of wACSFs leads to significantly better generalization performance in the machine learning potential than the large set of conventional ACSFs. Furthermore, we show that the intrinsic parameters of the descriptors can in principle be optimized with a genetic algorithm in a highly automated manner. For the wACSFs employed here, we find however that using a simple empirical parametrization scheme is sufficient in order to obtain HDNNPs with high accuracy.

References in corpus (6)

• ANI-1: An extensible neural network potential with DFT accuracy at force field computational cost
• Efficient and Accurate Machine-Learning Interpolation of Atomic Energies in Compositions with Many Species
• Resolving transition metal chemical space: feature selection for machine learning and structure-property relationships
• Understanding molecular representations in machine learning: The role of uniqueness and target similarity
• Pure density functional for strong correlations and the thermodynamic limit from machine learning
• Comparing the Accuracy of High-Dimensional Neural Network Potentials and the Systematic Molecular Fragmentation Method: A Benchmark Study for all-trans Alkanes

Cited by in corpus (75)

• Machine Learning Force Fields
• Combining Machine Learning and Computational Chemistry for Predictive Insights Into Chemical Systems
• DScribe: Library of Descriptors for Machine Learning in Materials Science
• Data-driven materials science: status, challenges and perspectives
• DeePMD-kit v2: A software package for Deep Potential models
• Neuroevolution machine learning potentials: Combining high accuracy and low cost in atomistic simulations and application to heat transport
• Physics-inspired structural representations for molecules and materials
• GPUMD: A package for constructing accurate machine-learned potentials and performing highly efficient atomistic simulations
• SchNetPack: A Deep Learning Toolbox For Atomistic Systems
• Machine learning for electronically excited states of molecules
• SpookyNet: Learning Force Fields with Electronic Degrees of Freedom and Nonlocal Effects
• FCHL revisited: faster and more accurate quantum machine learning
• Automatic Selection of Atomic Fingerprints and Reference Configurations for Machine-Learning Potentials
• Embedded Atom Neural Network Potentials: Efficient and Accurate Machine Learning with a Physically Inspired Representation
• A practical guide to machine learning interatomic potentials -- Status and future
• Atom-Density Representations for Machine Learning
• Machine Learning in QM/MM Molecular Dynamics Simulations of Condensed-Phase Systems
• Representations of molecules and materials for interpolation of quantum-mechanical simulations via machine learning
• Improving the accuracy of the neuroevolution machine learning potential for multi-component systems
• Feature Optimization for Atomistic Machine Learning Yields A Data-Driven Construction of the Periodic Table of the Elements
• Neural Network Potentials for Chemistry: Concepts, Applications and Prospects
• DPA-2: a large atomic model as a multi-task learner
• Advances of Machine Learning in Materials Science: Ideas and Techniques
• Guest Editorial: Special Topic on Data-enabled Theoretical Chemistry
• Physically Motivated Recursively Embedded Atom Neural Networks: Incorporating Local Completeness and Nonlocality
• Strategies for the Construction of Machine-Learning Potentials for Accurate and Efficient Atomic-Scale Simulations
• Solid Harmonic Wavelet Scattering for Predictions of Molecule Properties
• Tutorial: How to Train a Neural Network Potential
• Recursive evaluation and iterative contraction of $N$-body equivariant features
• PiNN: A Python Library for Building Atomic Neural Networks of Molecules and Materials
• Physically inspired deep learning of molecular excitations and photoemission spectra
• Machine learning and excited-state molecular dynamics
• The Evolution of Machine Learning Potentials for Molecules, Reactions and Materials
• PANNA: Properties from Artificial Neural Network Architectures
• An assessment of the structural resolution of various fingerprints commonly used in machine learning
• Scalable Hybrid Deep Neural Networks/Polarizable Potentials Biomolecular Simulations including long-range effects
• Lifelong Machine Learning Potentials
• Machine learning enhanced global optimization by clustering local environments to enable bundled atomic energies
• Atomic energy mapping of neural network potential
• Accelerating Atomistic Simulations with Piecewise Machine Learned Ab Initio Potentials at Classical Force Field-like Cost
• Tensor-reduced atomic density representations
• A New Kind of Atlas of Zeolite Building Blocks
• Genarris 2.0: A Random Structure Generator for Molecular Crystals
• PyXtal FF: a Python Library for Automated Force Field Generation
• Deep Neural Network Computes Electron Densities and Energies of a Large Set of Organic Molecules Faster than Density Functional Theory (DFT)
• E(n)-Equivariant Cartesian Tensor Passing Potential
• Learning from the Density to Correct Total Energy and Forces in First Principle Simulations
• Manifolds of quasi-constant SOAP and ACSF fingerprints and the resulting failure to machine learn four-body interactions
• High Pressure and Temperature Neural Network Reactive Force Field for Energetic Materials
• Accurate Deep Learning-aided Density-free Strategy for Many-Body Dispersion-corrected Density Functional Theory
• Quantum neural networks force fields generation
• Kernel based quantum machine learning at record rate : Many-body distribution functionals as compact representations
• Atomic structure optimization with machine-learning enabled interpolation between chemical elements
• An \textit{ab initio} study of shock-compressed copper
• Modeling of many-body interactions between elastic spheres through symmetry functions
• A transferable artificial neural network model for atomic forces in nanoparticles
• GPU-Accelerated Approximate Kernel Method for Quantum Machine Learning
• Neural Networks Potential from the Bispectrum Component: A Case Study on Crystalline Silicon
• Anisotropic molecular coarse-graining by force and torque matching with neural networks
• Mean-Field Density Matrix Decompositions
• Quantum Extreme Learning of molecular potential energy surfaces and force fields
• Compressing physical properties of atomic species for improving predictive chemistry
• Machine Learning Enhanced Calculation of Quantum-Classical Binding Free Energies
• Developing Potential Energy Surfaces for Graphene-based 2D-3D Interfaces from Modified High Dimensional Neural Networks for Applications in Energy Storage
• Fortnet, a software package for training Behler-Parrinello neural networks
• EOSnet: Embedded Overlap Structures for Graph Neural Networks in Predicting Material Properties
• Lifelong Machine Learning Potentials for Chemical Reaction Network Explorations
• Physical machine learning outperforms "human learning" in Quantum Chemistry
• The Homunculus Brain and Categorical Logic
• Efficient interpolation of molecular properties across chemical compound space with low-dimensional descriptors
• Efficient Data Selection Methods for the Development of Machine Learned Potentials
• Compressing local atomic neighbourhood descriptors
• Bond type restricted radial distribution functions for accurate machine learning prediction of atomization energies
• REANN: A PyTorch-based End-to-End Multi-functional Deep Neural Network Package for Molecular, Reactive and Periodic Systems
• Active Learning of a Neural Network Potential for Large-Scale Atomistic Simulations of Polymer Electrolyte Membranes