Representing molecule-surface interactions with symmetry-adapted neural networks

Abstract

The accurate description of molecule-surface interactions requires a detailed knowledge of the underlying potential-energy surface (PES). Recently, neural networks (NNs) have been shown to be an efficient technique to accurately interpolate the PES information provided for a set of molecular configurations, e.g. by first-principles calculations. Here, we further develop this approach by building the NN on a new type of symmetry functions, which allows to take the symmetry of the surface exactly into account. The accuracy and efficiency of such symmetry-adapted NNs is illustrated by the application to a six-dimensional PES describing the interaction of oxygen molecules with the Al(111) surface.

13 pages including 8 figures; related publications can be found at http://www.fhi-berlin.mpg.de/th/th.html

References in corpus (5)

• Six-Dimensional Quantum Dynamics of Adsorption and Desorption of H_2 at Pd(100): Steering and Steric Effects
• Dissociation of O2 at Al(111): The Role of Spin Selection Rules
• Ab initio Quantum and ab initio Molecular Dynamics of the Dissociative Adsorption of Hydrogen on Pd(100)
• Non-Adiabatic Potential-Energy Surfaces by Constrained Density-Functional Theory
• Ab initio calculation of the potential energy surface for the dissociation of H_2 on the sulfur covered Pd(100) surface

Cited by in corpus (31)

• Machine Learning of Accurate Energy-Conserving Molecular Force Fields
• Physics-inspired structural representations for molecules and materials
• Machine learning for electronically excited states of molecules
• A neural network interatomic potential for the phase change material GeTe
• Efficient Bayesian Inference of Atomistic Structure in Complex Functional Materials
• Non-adiabatic Effects in the Dissociation of Oxygen Molecules at the Al(111) Surface
• Ab initio quality neural-network potential for sodium
• Molecular Force Fields with Gradient-Domain Machine Learning: Construction and Application to Dynamics of Small Molecules with Coupled Cluster Forces
• Perspective: Energy Landscapes for Machine Learning
• Fingerprints for spin-selection rules in the interaction dynamics of O2 at Al(111)
• First-principles interatomic potentials for ten elemental metals via compressed sensing
• Signatures of nonadiabatic O2 dissociation at Al(111): First-principles fewest-switches study
• The Evolution of Machine Learning Potentials for Molecules, Reactions and Materials
• Machine learning and excited-state molecular dynamics
• Non-adiabatic effects during the dissociative adsorption of O2 at Ag(111)? A first-principles divide and conquer study
• Molecular Force Fields with Gradient-Domain Machine Learning (GDML): Comparison and Synergies with Classical Force Fields
• Challenges for Machine Learning Force Fields in Reproducing Potential Energy Surfaces of Flexible Molecules
• A deep neural network for molecular wave functions in quasi-atomic minimal basis representation
• Closing the gap between theory and experiment for lithium manganese oxide spinels using a high-dimensional neural network potential
• Nuclear-Electronic All-Particle Density Matrix Renormalization Group
• On the role of physisorption states in molecular scattering: A semi-local density-functional theory study on O2/Ag(111)
• Machine Learning and Data-Driven Methods in Computational Surface and Interface Science
• Quantum Proton Effects from Density Matrix Renormalization Group Calculations
• Machine Learning for Exploring Small Polaron Configurational Space
• Ready, set and no action: A static perspective on potential energy surfaces commonly used in gas-surface dynamics
• Instanton rate constant calculations using interpolated potential energy surfaces in non-redundant, rotationally and translationally invariant coordinates
• Construction of Machine Learned Force Fields with Quantum Chemical Accuracy: Applications and Chemical Insights
• Iterative charge equilibration for fourth-generation high-dimensional neural network potentials
• Compressing physical properties of atomic species for improving predictive chemistry
• Reliable computational prediction of supramolecular ordering of complex molecules under electrochemical conditions
• Computational design of metal-supported molecular switches: Transient ion formation during light- and electron-induced isomerisation of azobenzene