Accelerating the search for global minima on potential energy surfaces using machine learning

2016 ◽  
Vol 145 (15) ◽  
pp. 154106 ◽  
Author(s):  
S. F. Carr ◽  
R. Garnett ◽  
C. S. Lo
Keyword(s):  
Machine Learning ◽  
Potential Energy ◽  
Potential Energy Surfaces ◽  
Global Minima ◽  
Energy Surfaces

scholarly journals Planar Tetracoordinate Carbons in Allene-Type Structures

2021 ◽  
Author(s):  
Meng-hui Wang ◽  
Mesias Orozco-Ic ◽  
luis leyva-parra ◽  
William Tiznado ◽  
jorge barroso ◽  
...  
Keyword(s):  
Potential Energy ◽  
Potential Energy Surfaces ◽  
Electrostatic Interactions ◽  
Magnetic Response ◽  
Global Minima ◽  
Covalent Bonds ◽  
Bonding Analysis ◽  
Localization Strategy ◽  
Energy Surfaces

<p>The exhaustive exploration of the potential energy surfaces of CE<sub>2</sub>M<sub>2</sub> (E = Si-Pb, M = Li, and Na) revealed seven global minima containing a planar tetracoordinate carbon (ptC). The design was based on a π-localization strategy, resulting in a ptC embedded in a linear or a bent allene-type E=C=E motif. The magnetic response showed a σ-aromaticity in the bent E=C=E fragments. The bonding analysis indicated that the ptCs form C-E covalent bonds and C-M electrostatic interactions.<b></b></p>

scholarly journals Easy representation of multivariate functions with low-dimensional terms via Gaussian process regression kernel design: applications to machine learning of potential energy surfaces and kinetic energy densities from sparse data

2022 ◽  
Author(s):  
Sergei Manzhos ◽  
Eita Sasaki ◽  
Manabu Ihara
Keyword(s):  
Machine Learning ◽  
Kinetic Energy ◽  
Potential Energy ◽  
Gaussian Process ◽  
Potential Energy Surfaces ◽  
Gaussian Process Regression ◽  
Multivariate Functions ◽  
Low Dimensional ◽  
Energy Surfaces ◽  
Kernel Design

Abstract We show that Gaussian process regression (GPR) allows representing multivariate functions with low-dimensional terms via kernel design. When using a kernel built with HDMR (High-dimensional model representation), one obtains a similar type of representation as the previously proposed HDMR-GPR scheme while being faster and simpler to use. We tested the approach on cases where highly accurate machine learning is required from sparse data by fitting potential energy surfaces and kinetic energy densities.

scholarly journals Machine learning for potential energy surfaces: An extensive database and assessment of methods

2019 ◽  
Vol 150 (24) ◽  
pp. 244113 ◽  
Author(s):  
Gunnar Schmitz ◽  
Ian Heide Godtliebsen ◽  
Ove Christiansen
Keyword(s):  
Machine Learning ◽  
Potential Energy ◽  
Potential Energy Surfaces ◽  
Energy Surfaces

scholarly journals High-dimensional potential energy surfaces for molecular simulations: from empiricism to machine learning

2020 ◽  
Vol 1 (1) ◽  
pp. 013001 ◽  
Author(s):  
Oliver T Unke ◽  
Debasish Koner ◽  
Sarbani Patra ◽  
Silvan Käser ◽  
Markus Meuwly
Keyword(s):  
Machine Learning ◽  
Potential Energy ◽  
Potential Energy Surfaces ◽  
Molecular Simulations ◽  
High Dimensional ◽  
Energy Surfaces

scholarly journals van der Waals potential energy surfaces from the exchange-hole dipole moment dispersion model

2016 ◽  
Vol 94 (12) ◽  
pp. 1049-1056 ◽  
Author(s):  
Joseph B. Dizon ◽  
Erin R. Johnson
Keyword(s):  
Dipole Moment ◽  
Potential Energy ◽  
Potential Energy Surfaces ◽  
Reference Data ◽  
Dispersion Model ◽  
Binding Energies ◽  
Coupled Cluster ◽  
Global Minima ◽  
Hybrid Functional ◽  
Energy Surfaces

The potential energy surfaces (PESs) of 28 simple van der Waals complexes, each consisting of a rare-gas (Rg) atom interacting with a linear molecule, are calculated using the exchange-hole dipole moment (XDM) dispersion model in conjunction with three base density functionals (HFPBE, PW86PBE, and a commensurate hybrid functional). Results are compared with literature coupled-cluster reference data. The quality of the computed PESs is assessed based on the positions of the global minima and the corresponding binding energies. Only the hybrid functional is found to provide generally reliable PESs. Dispersion-corrected HFPBE strongly underestimates the equilibrium intermolecular separations and predicts different global minima than the reference PESs for Rg–HCl, Rg–HBr, and two of the Rg–HCN complexes. Analysis of the binding-energy errors reveals that the performance of HFPBE degrades as the size of the Rg atoms increase down the group, while the performance of PW86PBE is significantly worse for strongly-polar molecules. PW86PBE, and to a lesser extent the hybrid, strongly overbind Kr–HF due to charge-transfer error. Despite this, the XDM-corrected hybrid functional displays the best overall error statistics and provides binding energies to within ca. 10 cm–1 of the coupled-cluster reference data at a greatly reduced computational cost.

scholarly journals DIFFUSION MONTE CARLO USING MACHINE LEARNING POTENTIAL ENERGY SURFACES

2021 ◽  
Author(s):  
Ryan DiRisio ◽  
Anne McCoy ◽  
Fenris Lu ◽  
Jacob Finney ◽  
Mark Boyer
Keyword(s):  
Machine Learning ◽  
Monte Carlo ◽  
Potential Energy ◽  
Potential Energy Surfaces ◽  
Learning Potential ◽  
Diffusion Monte Carlo ◽  
Energy Surfaces

scholarly journals Planar Tetracoordinate Carbons in Allene-Type Structures

2021 ◽  
Author(s):  
Meng-hui Wang ◽  
Mesias Orozco-Ic ◽  
luis leyva-parra ◽  
William Tiznado ◽  
jorge barroso ◽  
...  
Keyword(s):  
Potential Energy ◽  
Potential Energy Surfaces ◽  
Electrostatic Interactions ◽  
Magnetic Response ◽  
Global Minima ◽  
Covalent Bonds ◽  
Bonding Analysis ◽  
Localization Strategy ◽  
Energy Surfaces

<p>The exhaustive exploration of the potential energy surfaces of CE<sub>2</sub>M<sub>2</sub> (E = Si-Pb, M = Li, and Na) revealed seven global minima containing a planar tetracoordinate carbon (ptC). The design was based on a π-localization strategy, resulting in a ptC embedded in a linear or a bent allene-type E=C=E motif. The magnetic response showed a σ-aromaticity in the bent E=C=E fragments. The bonding analysis indicated that the ptCs form C-E covalent bonds and C-M electrostatic interactions.<b></b></p>

Progress towards machine learning reaction rate constants

2021 ◽  
Author(s):  
Evan Komp ◽  
Nida Janulaitis ◽  
Stephanie Valleau
Keyword(s):  
Machine Learning ◽  
Potential Energy ◽  
Rate Constant ◽  
Rate Constants ◽  
Potential Energy Surfaces ◽  
Reaction Rate ◽  
Reaction Rate Constant ◽  
Reaction Rate Constants ◽  
The Cost ◽  
Energy Surfaces

Quantum and classical reaction rate constant calculations come at the cost of exploring potential energy surfaces. Due to the “curse of dimensionality”, their evaluation quickly becomes unfeasible as the system...

Sign in / Sign up

Export Citation Format

Share Document