DP-GEN: A concurrent learning platform for the generation of reliable deep learning based potential energy models

Abstract

In recent years, promising deep learning based interatomic potential energy surface (PES) models have been proposed that can potentially allow us to perform molecular dynamics simulations for large scale systems with quantum accuracy. However, making these models truly reliable and practically useful is still a very non-trivial task. A key component in this task is the generation of datasets used in model training. In this paper, we introduce the Deep Potential GENerator (DP-GEN), an open-source software platform that implements the recently proposed "on-the-fly" learning procedure [Phys. Rev. Materials 3, 023804] and is capable of generating uniformly accurate deep learning based PES models in a way that minimizes human intervention and the computational cost for data generation and model training. DP-GEN automatically and iteratively performs three steps: exploration, labeling, and training. It supports various popular packages for these three steps: LAMMPS for exploration, Quantum Espresso, VASP, CP2K, etc. for labeling, and DeePMD-kit for training. It also allows automatic job submission and result collection on different types of machines, such as high performance clusters and cloud machines, and is adaptive to different job management tools, including Slurm, PBS, and LSF. As a concrete example, we illustrate the details of the process for generating a general-purpose PES model for Cu using DP-GEN.

29 pages, 6 figures

References in corpus (12)

• TensorFlow: Large-Scale Machine Learning on Heterogeneous Distributed Systems
• Gaussian Approximation Potentials: the accuracy of quantum mechanics, without the electrons
• Deep Potential Molecular Dynamics: a scalable model with the accuracy of quantum mechanics
• DeePMD-kit: A deep learning package for many-body potential energy representation and molecular dynamics
• Less is more: sampling chemical space with active learning
• Active learning of linearly parametrized interatomic potentials
• AiiDA: Automated Interactive Infrastructure and Database for Computational Science
• Error estimates for solid-state density-functional theory predictions: an overview by means of the ground-state elemental crystals
• Active Learning of Uniformly Accurate Inter-atomic Potentials for Materials Simulation
• SchNetPack: A Deep Learning Toolbox For Atomistic Systems
• Deep Potential: a general representation of a many-body potential energy surface
• i-PI: A Python interface for ab initio path integral molecular dynamics simulations

Cited by in corpus (96)

• DeePMD-kit v2: A software package for Deep Potential models
• The Phase Diagram of a Deep Potential Water Model
• Neural Network Based in Silico Simulation of Combustion Reactions
• Intelligent Computing: The Latest Advances, Challenges and Future
• 86 PFLOPS Deep Potential Molecular Dynamics simulation of 100 million atoms with ab initio accuracy
• A deep potential model with long-range electrostatic interactions
• Deep Potentials for Materials Science
• Algorithms for Solving High Dimensional PDEs: From Nonlinear Monte Carlo to Machine Learning
• Deep neural network for the dielectric response of insulators
• Raman Spectrum and Polarizability of Liquid Water from Deep Neural Networks
• DPA-2: a large atomic model as a multi-task learner
• Deep Potential generation scheme and simulation protocol for the Li10GeP2S12-type superionic conductors
• Dissolving salt is not equivalent to applying a pressure on water
• Advances of Machine Learning in Materials Science: Ideas and Techniques
• Phase equilibrium of water with hexagonal and cubic ice using the SCAN functional
• DeePMD-kit v3: A Multiple-Backend Framework for Machine Learning Potentials
• Automated discovery of a robust interatomic potential for aluminum
• Deep Learning of Accurate Force Field of Ferroelectric HfO${}_2$
• Structural Phase Transitions in SrTiO3 from Deep Potential Molecular Dynamics
• Accurate Deep Potential model for the Al-Cu-Mg alloy in the full concentration space
• Viscosity in water from first-principles and deep-neural-network simulations
• Liquid-Liquid Critical Point in Phosphorus
• Data-Driven Prediction of Complex Crystal Structures of Dense Lithium
• Phonons in magic-angle twisted bilayer graphene
• Modular development of deep potential for complex solid solutions
• Extending the limit of molecular dynamics with ab initio accuracy to 10 billion atoms
• Universal interatomic potential for perovskite oxides
• A Tungsten Deep Neural-Network Potential for Simulating Mechanical Property Degradation Under Fusion Service Environment
• Machine Learning Potential for Electrochemical Interfaces with Hybrid Representation of Dielectric Response
• Temperature- and vacancy-concentration-dependence of heat transport in Li${}_3$ClO from multi-method numerical simulations
• Understanding the Anomalous Diffusion of Water in Aqueous Electrolytes Using Machine Learned Potentials
• Pushing the limit of molecular dynamics with ab initio accuracy to 100 million atoms with machine learning
• ArcaNN: automated enhanced sampling generation of training sets for chemically reactive machine learning interatomic potentials
• Unconventional ferroelectric domain switching dynamics in CuInP2S6 from first principles
• Machine-learning interatomic potential for molecular dynamics simulation of ferroelectric KNbO3 perovskite
• A Deep Potential model for liquid-vapor equilibrium and cavitation rates of water
• Towards fully ab initio simulation of atmospheric aerosol nucleation
• XLB: A differentiable massively parallel lattice Boltzmann library in Python
• Integrating Machine Learning with Physics-Based Modeling
• Innate Dynamics and Identity Crisis of a Metal Surface Unveiled by Machine Learning of Atomic Environments
• Theoretical evidence of H-He demixing under Jupiter and Saturn conditions
• Accelerating the prediction of inorganic surfaces with machine learning interatomic potentials
• Origin of negative thermal expansion and pressure induced amorphization in zirconium tungstate from machine-learning potential
• Lattice dynamics and elastic properties of alpha-U at high-temperature and high-pressure by machine learning potential simulations
• Finite-temperature properties of antiferroelectric perovskite $\mathrm{PbZr}{\mathrm{O}}_3$ from deep learning interatomic potential
• Modeling the High-Pressure Solid and Liquid Phases of Tin from Deep Potentials with ab initio Accuracy
• A "Magnetic" Machine Learning Interatomic Potential for Nickel
• Large-scale Atomistic Simulation of Quantum Effects in SrTiO${}_3$ from First Principles
• Structure and polymerization of liquid sulfur across the $\lambda$-transition
• DeePKS+ABACUS as a Bridge between Expensive Quantum Mechanical Models and Machine Learning Potentials
• Generator of Neural Network Potential for Molecular Dynamics: Constructing Robust and Accurate Potentials with Active Learning for Nanosecond-scale Simulations
• Real-Space Visualization of Frequency-Dependent Anisotropy of Atomic Vibrations
• Chirality-Dependent Kinetics of Single-Walled Carbon Nanotubes from Machine-Learning Force Fields
• Training Algorithm Matters for the Performance of Neural Network Potential: A Case Study of Adam and the Kalman Filter Optimizers
• Validation Workflow for Machine Learning Interatomic Potentials for Complex Ceramics
• How Graph Neural Network Interatomic Potentials Extrapolate: Role of the Message-Passing Algorithm
• Machine learning based nonlocal kinetic energy density functional for simple metals and alloys
• Structure and mechanical properties of monolayer amorphous carbon and boron nitride
• Unraveling Antagonistic Collision-Controlled Reactivity in Energetic Molecular Perovskites with Deep Potential Molecular Dynamics
• Thermal disorder and phonon softening in the ferroelectric phase transition of lead titanate
• High-Throughput Condensed-Phase Hybrid Density Functional Theory for Large-Scale Finite-Gap Systems: The SeA Approach
• DPmoire: A tool for constructing accurate machine learning force fields in moiré systems
• Topological phase transitions in perovskite superlattices driven by temperature, electric field, and doping
• Probing the state of hydrogen in $\delta$-AlOOH at mantle conditions with machine learning potential
• Non-Equilibrium Nature of Fracture Determines the Crack Paths
• Glassy Dynamics from First-Principles Simulations
• Local structure, thermodynamics, and melting curve of boron phosphide at high pressures by deep learning-driven ab initio simulations
• Multi-channel machine learning based nonlocal kinetic energy density functional for semiconductors
• Pre-training, fine-tuning, and distillation (PFD): Automatically generating machine learning force fields from universal models
• KPROJ: A Program for Unfolding Electronic and Phononic Bands
• Using metadynamics to build neural network potentials for reactive events: the case of urea decomposition in water
• Aqueous Solution Chemistry In Silico and the Role of Data Driven Approaches
• On the Quality of Uncertainty Estimates from Neural Network Potential Ensembles
• Deep Learning Inter-atomic Potential for Thermal and Phonon Behaviour of Silicon Carbide with Quantum Accuracy
• Modified Steinberg-Guinan elasticity model to describe softening-hardening dual anomaly in vanadium
• A Planning-and-Exploring Approach to Extreme-Mechanics Force Fields
• Slowly Quenched, High Pressure Glassy B${}_2$O${}_3$ at DFT Accuracy
• Atomistic simulations of thermodynamic properties of liquid gallium from first principles
• Structure and solidification of the (Fe0.75B0.15Si0.1)100-xTax (x=0-2) melts: experiment and machine learning
• Transferability of the chemical bond-based machine learning model for dipole moment: the GHz to THz dielectric properties of liquid propylene glycol and polypropylene glycol
• Strength of 2D Glasses Explored by Machine-Learning Force Fields
• Machine-Learned Potentials for Solvation Modeling
• Learning atomic forces from uncertainty-calibrated adversarial attacks
• Deep variational free energy prediction of dense hydrogen solid at 1200K
• Self-healing mechanism of lithium in lithium metal batteries
• Considerations in the use of ML interaction potentials for free energy calculations
• Ab Initio bulk free energy surface of proper ferroelectrics
• Training Data Set Refinement for the Machine Learning Potential of Li-Si Alloys via Structural Similarity Analysis
• Accurate force field of two-dimensional ferroelectrics from deep learning
• High-pressure melting and elastic behavior of vanadium and niobium based on ab initio and machine learning molecular dynamics
• Active Learning of a Neural Network Potential for Large-Scale Atomistic Simulations of Polymer Electrolyte Membranes
• Intrinsic structure of relaxor ferroelectrics from first principles
• Location of the liquid-vapor critical point in aluminum
• Machine learning potential as a guide for eutectic in ultra-refractory multicomponent ceramics
• Deformation mechanisms and compressive response of NbTaTiZr alloy via machine learning potentials
• The ground state of CuInP${}_2$S${}_6$ thin films: A study of the deep potential method