Algorithmic Differentiation for Automated Modeling of Machine Learned Force Fields

Abstract

Reconstructing force fields (FFs) from atomistic simulation data is a challenge since accurate data can be highly expensive. Here, machine learning (ML) models can help to be data economic as they can be successfully constrained using the underlying symmetry and conservation laws of physics. However, so far, every descriptor newly proposed for an ML model has required a cumbersome and mathematically tedious remodeling. We therefore propose using modern techniques from algorithmic differentiation within the ML modeling process -- effectively enabling the usage of novel descriptors or models fully automatically at an order of magnitude higher computational efficiency. This paradigmatic approach enables not only a versatile usage of novel representations and the efficient computation of larger systems -- all of high value to the FF community -- but also the simple inclusion of further physical knowledge such as higher-order information (e.g. Hessians, more complex partial differential equations constraints etc.), even beyond the presented FF domain.

arXiv comment17 pages, 4 figures

References in corpus (27)

• Gaussian Approximation Potentials: the accuracy of quantum mechanics, without the electrons
• Machine Learning for Fluid Mechanics
• Fast and Accurate Modeling of Molecular Atomization Energies with Machine Learning
• SchNet - a deep learning architecture for molecules and materials
• Machine Learning of Accurate Energy-Conserving Molecular Force Fields
• Searching for Exotic Particles in High-Energy Physics with Deep Learning
• PhysNet: A Neural Network for Predicting Energies, Forces, Dipole Moments and Partial Charges
• Machine learning for molecular simulation
• Combining Machine Learning and Computational Chemistry for Predictive Insights Into Chemical Systems
• Towards Exact Molecular Dynamics Simulations with Machine-Learned Force Fields
• Machine Learning of Linear Differential Equations using Gaussian Processes
• Tensor field networks: Rotation- and translation-equivariant neural networks for 3D point clouds
• VAMPnets: Deep learning of molecular kinetics
• Artificial Intelligence and Statistics
• SpookyNet: Learning Force Fields with Electronic Degrees of Freedom and Nonlocal Effects
• FCHL revisited: faster and more accurate quantum machine learning
• Physically-informed artificial neural networks for atomistic modeling of materials
• Alchemical and structural distribution based representation for improved QML
• sGDML: Constructing Accurate and Data Efficient Molecular Force Fields Using Machine Learning
• Flow-based generative models for Markov chain Monte Carlo in lattice field theory
• Incorporating long-range physics in atomic-scale machine learning
• Equivariant flow-based sampling for lattice gauge theory
• Estimation of Thermodynamic Observables in Lattice Field Theories with Deep Generative Models
• Operators in Machine Learning: Response Properties in Chemical Space
• Gauge equivariant neural networks for quantum lattice gauge theories
• Variational estimation of the drift for stochastic differential equations from the empirical density
• Accurate global machine learning force fields for molecules with hundreds of atoms

Cited by in corpus (6)

• Stress and heat flux via automatic differentiation
• Redundancy Resolution at Position Level
• Machine-Learned Potentials for Solvation Modeling
• Scaling up machine learning-based chemical plant simulation: A method for fine-tuning a model to induce stable fixed points
• Reconstructing Kernel-based Machine Learning Force Fields with Super-linear Convergence
• Algorithmic differentiation for plane-wave DFT: materials design, error control and learning model parameters