Skip to main content
OSTI.GOVU.S. Department of Energy Office of Scientific and Technical Information
U.S. Department of Energy
Office of Scientific and Technical Information
 

PANNA: Properties from Artificial Neural Network Architectures

Journal Article · · Computer Physics Communications
 [1];  [2];  [2];  [2]
  1. International School for Advanced Studies (SISSA), Trieste (Italy); International School for Advanced Studies (SISSA), Trieste (Italy)
  2. International School for Advanced Studies (SISSA), Trieste (Italy)
+ Show Author Affiliations

We report prediction of material properties from first principles is often a computationally expensive task. Recently, artificial neural networks and other machine learning approaches have been successfully employed to obtain accurate models at a low computational cost by leveraging existing example data. Here, we present a software package “Properties from Artificial Neural Network Architectures” (PANNA) that provides a comprehensive toolkit for creating neural network models for atomistic systems following the Behler–Parrinello topology. Besides the core routines for neural network training, it includes data parser, descriptor builder for Behler–Parrinello class of symmetry functions and force-field generator suitable for integration within molecular dynamics packages. PANNA offers a variety of activation and cost functions, regularization methods, as well as the possibility of using fully-connected networks with custom size for each atomic species. PANNA benefits from the optimization and hardware-flexibility of the underlying TensorFlow engine which allows it to be used on multiple CPU/GPU/TPU systems, making it possible to develop and optimize neural network models based on large datasets.

Research Organization:
Harvard Univ., Cambridge, MA (United States)
Sponsoring Organization:
USDOE Office of Science (SC), Basic Energy Sciences (BES); European Union’s Horizon 2020; MIT International Science and Technology Initiatives; MIT International Science and Technology Initiatives
Grant/Contract Number:
SC0019300
OSTI ID:
1853030
Alternate ID(s):
OSTI ID: 1775673
Journal Information:
Computer Physics Communications, Journal Name: Computer Physics Communications Journal Issue: C Vol. 256; ISSN 0010-4655
Publisher:
ElsevierCopyright Statement
Country of Publication:
United States
Language:
English

References (66)

How Chemical Composition Alone Can Predict Vibrational Free Energies and Entropies of Solids journal July 2017
PhysNet: A Neural Network for Predicting Energies, Forces, Dipole Moments, and Partial Charges journal April 2019
Fast and Accurate Molecular Property Prediction: Learning Atomic Interactions and Potentials with Neural Networks journal August 2018
Understanding the Composition and Activity of Electrocatalytic Nanoalloys in Aqueous Solvents: A Combination of DFT and Accurate Neural Network Potentials journal April 2014
Bypassing the Kohn-Sham equations with machine learning journal October 2017
SchNet – A deep learning architecture for molecules and materials journal June 2018
Advanced capabilities for materials modelling with Quantum ESPRESSO journal October 2017
Accurate interatomic force fields via machine learning with covariant kernels journal June 2017
Neural-network-enhanced evolutionary algorithm applied to supported metal nanoparticles journal May 2018
Molecular Dynamics with On-the-Fly Machine Learning of Quantum-Mechanical Forces journal March 2015
Universal fragment descriptors for predicting properties of inorganic crystals text January 2017
Development of a machine learning potential for graphene text January 2018
Transferable Molecular Charge Assignment Using Deep Neural Networks preprint January 2018
Machine Learning Predictions of Molecular Properties: Accurate Many-Body Potentials and Nonlocality in Chemical Space text January 2015
Fast Parallel Algorithms for Short-Range Molecular Dynamics journal March 1995
The potential of atomistic simulations and the knowledgebase of interatomic models journal July 2011
XCrySDen—a new program for displaying crystalline structures and electron densities journal June 1999
An implementation of artificial neural-network potentials for atomistic materials simulations: Performance for TiO2 journal March 2016
AFLOW-ML: A RESTful API for machine-learning predictions of materials properties journal September 2018
Amp: A modular approach to machine learning in atomistic simulations journal October 2016
DeePMD-kit: A deep learning package for many-body potential energy representation and molecular dynamics journal July 2018
sGDML: Constructing accurate and data efficient molecular force fields using machine learning journal July 2019
SIMPLE-NN: An efficient package for training and executing neural-network interatomic potentials journal September 2019
Machine-Learning-Assisted Accurate Band Gap Predictions of Functionalized MXene journal May 2018
Transferable Dynamic Molecular Charge Assignment Using Deep Neural Networks journal July 2018
Library-Based LAMMPS Implementation of High-Dimensional Neural Network Potentials journal January 2019
Machine Learning Predictions of Molecular Properties: Accurate Many-Body Potentials and Nonlocality in Chemical Space journal June 2015
Transferable Machine-Learning Model of the Electron Density journal December 2018
Virtual Exploration of the Chemical Universe up to 11 Atoms of C, N, O, F:  Assembly of 26.4 Million Structures (110.9 Million Stereoisomers) and Analysis for New Ring Systems, Stereochemistry, Physicochemical Properties, Compound Classes, and Drug Discovery journal January 2007
Theoretical Investigation of Optimized Structures of Thiolated Gold Cluster [Au 25 (SCH 3 ) 18 ] + journal January 2007
Universal fragment descriptors for predicting properties of inorganic crystals journal June 2017
Graph dynamical networks for unsupervised learning of atomic scale dynamics in materials journal June 2019
Coarse-graining auto-encoders for molecular dynamics journal December 2019
Two-dimensional materials from high-throughput computational exfoliation of experimentally known compounds journal February 2018
ANI-1, A data set of 20 million calculated off-equilibrium conformations for organic molecules journal December 2017
ANI-1: an extensible neural network potential with DFT accuracy at force field computational cost journal January 2017
Machine learning for the structure–energy–property landscapes of molecular crystals journal January 2018
The TensorMol-0.1 model chemistry: a neural network augmented with long-range physics journal January 2018
Atom-centered symmetry functions for constructing high-dimensional neural network potentials journal February 2011
Neural network models of potential energy surfaces journal September 1995
Research Update: The materials genome initiative: Data sharing and the impact of collaborative ab initio databases journal March 2016
wACSF—Weighted atom-centered symmetry functions as descriptors in machine learning potentials journal June 2018
Building machine learning force fields for nanoclusters journal June 2018
Machine learning of molecular electronic properties in chemical compound space journal September 2013
Inhomogeneous Electron Gas journal November 1964
Electronic Effects in the Elastic Constants of n -Type Silicon journal September 1967
Computer simulation of local order in condensed phases of silicon journal April 1985
Efficient iterative schemes for ab initio total-energy calculations using a plane-wave basis set journal October 1996
On representing chemical environments journal May 2013
How to represent crystal structures for machine learning: Towards fast prediction of electronic properties journal May 2014
Accurate interatomic force fields via machine learning with covariant kernels journal June 2017
Development of a machine learning potential for graphene journal February 2018
Implanted neural network potentials: Application to Li-Si alloys journal March 2018
Neural-network-enhanced evolutionary algorithm applied to supported metal nanoparticles journal May 2018
Density functional theory based neural network force fields from energy decompositions journal February 2019
Gaussian Approximation Potentials: The Accuracy of Quantum Mechanics, without the Electrons journal April 2010
Fast and Accurate Modeling of Molecular Atomization Energies with Machine Learning journal January 2012
Molecular Dynamics with On-the-Fly Machine Learning of Quantum-Mechanical Forces journal March 2015
Deep Potential Molecular Dynamics: A Scalable Model with the Accuracy of Quantum Mechanics journal April 2018
Crystal Graph Convolutional Neural Networks for an Accurate and Interpretable Prediction of Material Properties journal April 2018
Silicon Liquid Structure and Crystal Nucleation from Ab Initio Deep Metadynamics journal December 2018
Generalized Neural-Network Representation of High-Dimensional Potential-Energy Surfaces journal April 2007
XSEDE: Accelerating Scientific Discovery journal September 2014
Wavelet Scattering Regression of Quantum Chemical Energies journal January 2017
Two-dimensional materials from high-throughput computational exfoliation of experimentally known compounds dataset January 2019
Transferable Machine-Learning Model of the Electron Density dataset January 2019

Cited By (1)

JAX, M.D.: A Framework for Differentiable Physics text January 2019

Similar Records

Neural Network Interatomic Potentials
Technical Report · Tue Sep 01 00:00:00 EDT 2020 · OSTI ID:1678825
Livermore Big Artificial Neural Network Toolkit
Software · Fri Jul 01 00:00:00 EDT 2016 · OSTI ID:1271009
Livermore Big Artificial Neural Network Toolkit
Software · Sun Jul 17 20:00:00 EDT 2016 · OSTI ID:code-4424

Related Subjects

71 CLASSICAL AND QUANTUM MECHANICS
GENERAL PHYSICS
97 MATHEMATICS AND COMPUTING
force field
machine learning
molecular dynamics
neural network
potential energy surface