scholarly journals Assessing Conformer Energies using Electronic Structure and Machine Learning Methods

2020 ◽  
Author(s):  
Dakota Folmsbee ◽  
Geoffrey Hutchison
Keyword(s):  
Machine Learning ◽  
Electronic Structure ◽  
Density Functional ◽  
Large Scale ◽  
Single Point ◽  
Semiempirical Method ◽  
Coupled Cluster ◽  
Scale Evaluation ◽  
Machine Learning Methods ◽  
Electronic Structure Methods

We have performed a large-scale evaluation of current computational methods, including conventional small-molecule force fields, semiempirical, density functional, ab initio electronic structure methods, and current machine learning (ML) techniques to evaluate relative single-point energies. Using up to 10 local minima geometries across ~700 molecules, each optimized by B3LYP-D3BJ with single-point DLPNO-CCSD(T) triple-zeta energies, we consider over 6,500 single points to compare the correlation between different methods for both relative energies and ordered rankings of minima. We find promise from current ML methods and recommend methods at each tier of the accuracy-time tradeoff, particularly the recent GFN2 semiempirical method, the B97-3c density functional approximation, and RI-MP2 for accurate conformer energies. The ANI family of ML methods shows promise, particularly the ANI-1ccx variant trained in part on coupled-cluster energies. Multiple methods suggest continued improvements should be expected in both performance and accuracy.

scholarly journals Assessing Conformer Energies using Electronic Structure and Machine Learning Methods

2020 ◽  
Author(s):  
Dakota Folmsbee ◽  
Geoffrey Hutchison
Keyword(s):  
Machine Learning ◽  
Electronic Structure ◽  
Density Functional ◽  
Large Scale ◽  
Single Point ◽  
Semiempirical Method ◽  
Coupled Cluster ◽  
Scale Evaluation ◽  
Machine Learning Methods ◽  
Electronic Structure Methods

We have performed a large-scale evaluation of current computational methods, including conventional small-molecule force fields, semiempirical, density functional, ab initio electronic structure methods, and current machine learning (ML) techniques to evaluate relative single-point energies. Using up to 10 local minima geometries across ~700 molecules, each optimized by B3LYP-D3BJ with single-point DLPNO-CCSD(T) triple-zeta energies, we consider over 6,500 single points to compare the correlation between different methods for both relative energies and ordered rankings of minima. We find promise from current ML methods and recommend methods at each tier of the accuracy-time tradeoff, particularly the recent GFN2 semiempirical method, the B97-3c density functional approximation, and RI-MP2 for accurate conformer energies. The ANI family of ML methods shows promise, particularly the ANI-1ccx variant trained in part on coupled-cluster energies. Multiple methods suggest continued improvements should be expected in both performance and accuracy.

scholarly journals Assessing Conformer Energies using Electronic Structure and Machine Learning Methods

2020 ◽  
Author(s):  
Dakota Folmsbee ◽  
Geoffrey Hutchison
Keyword(s):  
Machine Learning ◽  
Electronic Structure ◽  
Density Functional ◽  
Large Scale ◽  
Single Point ◽  
Semiempirical Method ◽  
Coupled Cluster ◽  
Scale Evaluation ◽  
Machine Learning Methods ◽  
Electronic Structure Methods

We have performed a large-scale evaluation of current computational methods, including conventional small-molecule force fields, semiempirical, density functional, ab initio electronic structure methods, and current machine learning (ML) techniques to evaluate relative single-point energies. Using up to 10 local minima geometries across ~700 molecules, each optimized by B3LYP-D3BJ with single-point DLPNO-CCSD(T) triple-zeta energies, we consider over 6,500 single points to compare the correlation between different methods for both relative energies and ordered rankings of minima. We find promise from current ML methods and recommend methods at each tier of the accuracy-time tradeoff, particularly the recent GFN2 semiempirical method, the B97-3c density functional approximation, and RI-MP2 for accurate conformer energies. The ANI family of ML methods shows promise, particularly the ANI-1ccx variant trained in part on coupled-cluster energies. Multiple methods suggest continued improvements should be expected in both performance and accuracy.

scholarly journals A look at the density functional theory zoo with the advanced GMTKN55 database for general main group thermochemistry, kinetics and noncovalent interactions

2017 ◽  
Vol 19 (48) ◽  
pp. 32184-32215 ◽  
Author(s):  
Lars Goerigk ◽  
Andreas Hansen ◽  
Christoph Bauer ◽  
Stephan Ehrlich ◽  
Asim Najibi ◽  
...  
Keyword(s):  
Density Functional Theory ◽  
Electronic Structure ◽  
Density Functional ◽  
Noncovalent Interactions ◽  
Main Group ◽  
Density Functionals ◽  
Functional Theory ◽  
The Density Functional Theory ◽  
Benchmark Database ◽  
Electronic Structure Methods

We present the updated and extended GMTKN55 benchmark database for more accurate and extensive energetic evaluation of density functionals and other electronic structure methods with detailed guidelines for method users.

scholarly journals 3145 An Evaluation of Machine Learning and Traditional Statistical Methods for Discovery in Large-Scale Translational Data

2019 ◽  
Vol 3 (s1) ◽  
pp. 2-2
Author(s):  
Megan C Hollister ◽  
Jeffrey D. Blume
Keyword(s):  
Gene Expression ◽  
Machine Learning ◽  
Random Forest ◽  
Gene Expression Data ◽  
Large Scale ◽  
Second Generation ◽  
A Priori ◽  
Expression Data ◽  
P Values ◽  
Machine Learning Methods

OBJECTIVES/SPECIFIC AIMS: To examine and compare the claims in Bzdok, Altman, and Brzywinski under a broader set of conditions by using unbiased methods of comparison. To explore how to accurately use various machine learning and traditional statistical methods in large-scale translational research by estimating their accuracy statistics. Then we will identify the methods with the best performance characteristics. METHODS/STUDY POPULATION: We conducted a simulation study with a microarray of gene expression data. We maintained the original structure proposed by Bzdok, Altman, and Brzywinski. The structure for gene expression data includes a total of 40 genes from 20 people, in which 10 people are phenotype positive and 10 are phenotype negative. In order to find a statistical difference 25% of the genes were set to be dysregulated across phenotype. This dysregulation forced the positive and negative phenotypes to have different mean population expressions. Additional variance was included to simulate genetic variation across the population. We also allowed for within person correlation across genes, which was not done in the original simulations. The following methods were used to determine the number of dysregulated genes in simulated data set: unadjusted p-values, Benjamini-Hochberg adjusted p-values, Bonferroni adjusted p-values, random forest importance levels, neural net prediction weights, and second-generation p-values. RESULTS/ANTICIPATED RESULTS: Results vary depending on whether a pre-specified significance level is used or the top 10 ranked values are taken. When all methods are given the same prior information of 10 dysregulated genes, the Benjamini-Hochberg adjusted p-values and the second-generation p-values generally outperform all other methods. We were not able to reproduce or validate the finding that random forest importance levels via a machine learning algorithm outperform classical methods. Almost uniformly, the machine learning methods did not yield improved accuracy statistics and they depend heavily on the a priori chosen number of dysregulated genes. DISCUSSION/SIGNIFICANCE OF IMPACT: In this context, machine learning methods do not outperform standard methods. Because of this and their additional complexity, machine learning approaches would not be preferable. Of all the approaches the second-generation p-value appears to offer significant benefit for the cost of a priori defining a region of trivially null effect sizes. The choice of an analysis method for large-scale translational data is critical to the success of any statistical investigation, and our simulations clearly highlight the various tradeoffs among the available methods.

scholarly journals Relativistic Effects on the Equation-of-State of the Light Actinides

2005 ◽  
Vol 893 ◽  
Author(s):  
Alexander Landa ◽  
Per Söderlind
Keyword(s):  
Electronic Structure ◽  
Equation Of State ◽  
Density Functional ◽  
Relativistic Effects ◽  
Model Systems ◽  
Face Centered Cubic ◽  
Electronic Structure Methods ◽  
The Face ◽  
Face Centered ◽  
Structure Calculations

AbstractThe effect of the relativistic spin-orbit (SO)interaction on the bonding in the early actinides has been investigated by means of electronic-structure calculations. Specifically, the equation of state (EOS) for the face-centered cubic (fcc) model systems of these metals has been calculated from the first-principles density-functional (DFT) theory. Traditionally, the SO interaction in electronic-structure methods is implemented as a perturbation to the Hamiltonian that is solved for basis functions that explicitly do not depend on SO coupling. Here this approximation is shown to compare well with the fully relativistic Dirac treatment. It is further shown that SO coupling has a gradually increasing effect on the EOS as one proceeds through the actinides and the effect is diminished as density increases.

scholarly journals Introduction to first-principles electronic structure methods: Application to actinide materials

2006 ◽  
Vol 21 (12) ◽  
pp. 2979-2985 ◽  
Author(s):  
John E. Klepeis
Keyword(s):  
Electronic Structure ◽  
First Principles ◽  
Density Functional ◽  
Local Density ◽  
Background Information ◽  
Quantum Molecular Dynamics ◽  
Generalized Gradient ◽  
Sufficient Detail ◽  
Current Controversy ◽  
Electronic Structure Methods

This paper provides an introduction for non-experts to first-principles electronic structure methods that are widely used in condensed-matter physics. Particular emphasis is placed on giving the appropriate background information needed to better appreciate the use of these methods to study actinide and other materials. Specifically, the underlying theory is described in sufficient detail to enable an understanding of the relative strengths and weaknesses of the methods. In addition, the meaning of commonly used terminology is explained, including density functional theory (DFT), local density approximation (LDA), and generalized gradient approximation (GGA), as well as linear muffin-tin orbital (LMTO), linear augmented plane wave (LAPW), and pseudopotential methods. Methodologies that extend the basic theory to address specific limitations are also briefly discussed. Finally, a few illustrative applications are presented, including quantum molecular dynamics (QMD) simulations and studies of surfaces, impurities, and defects. The paper concludes by addressing the current controversy regarding magnetic calculations for actinide materials.

scholarly journals Large-Scale Electronic Structure Calculations of Vacancies in 4H-SiC Using the Heyd-Scuseria-Ernzerhof Screened Hybrid Density Functional

2011 ◽  
Vol 679-680 ◽  
pp. 261-264 ◽  
Author(s):  
Tamas Hornos ◽  
Adam Gali ◽  
Bengt Gunnar Svensson
Keyword(s):  
Electronic Structure ◽  
Density Functional ◽  
Large Scale ◽  
Electronic Structure Calculations ◽  
Carbon Vacancy ◽  
Correction Scheme ◽  
Hybrid Density Functional ◽  
Structure Calculations

Large-scale and gap error free calculations of the electronic structure of vacancies in 4H-SiC have been carried out using a hybrid density functional (HSE06) and an accurate charge correction scheme. Based on the results the carbon vacancy is proposed to be responsible for the Z1/2 and EH6/7 DLTS centers.

scholarly journals Assessment of DFT Approaches in Noble Gas Clathrate-like Clusters: Stability and Thermodynamics

2021 ◽  
Author(s):  
Raquel Yanes-Rodríguez ◽  
Rita Prosmiti
Keyword(s):  
Electronic Structure ◽  
Density Functional ◽  
Noble Gas ◽  
Natural Orbital ◽  
Electronic Structure Methods ◽  
Local Pair

We have assessed the performance and accuracy of different wavefunction-based electronic structure methods, such as DFMP2 and domain-based local pair-natural orbital (DLPNO-CCSD(T)), as well as a variety of density functional...

Sign in / Sign up

Export Citation Format

Share Document