Atomistic simulations of corrosion related species in nano-cracks

Kallikragas, Dimitrios T.; Svishchev, Igor M.

doi:10.1016/j.corsci.2018.02.056

Cited by 5 publications

(2 citation statements)

References 45 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…With the enormous rise in computational power and the number of molecular simulation methods in the past decades, atomistic modeling is increasingly becoming the method of choice. [1][2][3][4][5][6][7][8][9] Applications range from the study and prevention of corrosion [10][11][12] to protein folding, 13 unfolding, 14 and self-assembly. 15 For many applications, machine learning force fields (MLFFs) are becoming the method of choice, as they can potentially reproduce any functional form of interatomic and intermolecular interactions, leading to reliable descriptions of potential energy surfaces (PESs) of arbitrary complexity.…”

Section: Introductionmentioning

confidence: 99%

Improving molecular force fields across configurational space by combining supervised and unsupervised machine learning

Fonseca

Poltavsky

Vassilev-Galindo

et al. 2021

The Journal of Chemical Physics

View full text Add to dashboard Cite

The training set of atomic configurations is key to the performance of any Machine Learning Force Field (MLFF) and, as such, the training set selection determines the applicability of the MLFF model for predictive molecular simulations. However, most atomistic reference datasets are inhomogeneously distributed across configurational space (CS), and thus, choosing the training set randomly or according to the probability distribution of the data leads to models whose accuracy is mainly defined by the most common close-to-equilibrium configurations in the reference data. In this work, we combine unsupervised and supervised ML methods to bypass the inherent bias of the data for common configurations, effectively widening the applicability range of the MLFF to the fullest capabilities of the dataset. To achieve this goal, we first cluster the CS into subregions similar in terms of geometry and energetics. We iteratively test a given MLFF performance on each subregion and fill the training set of the model with the representatives of the most inaccurate parts of the CS. The proposed approach has been applied to a set of small organic molecules and alanine tetrapeptide, demonstrating an up to twofold decrease in the root mean squared errors for force predictions on non-equilibrium geometries of these molecules. Furthermore, our ML models demonstrate superior stability over the default training approaches, allowing reliable study of processes involving highly out-of-equilibrium molecular configurations. These results hold for both kernel-based methods (sGDML and GAP/SOAP models) and deep neural networks (SchNet model).

show abstract

Section: Introductionmentioning

confidence: 99%

Improving molecular force fields across configurational space by combining supervised and unsupervised machine learning

Fonseca

Poltavsky

Vassilev-Galindo

et al. 2021

The Journal of Chemical Physics

View full text Add to dashboard Cite

show abstract

“…With the enormous rise in computational power and molecular simulation methods in the last decades, atomistic modeling is increasingly becoming the method of choice [1][2][3][4][5][6][7][8][9] . Applications range from the study and prevention of corrosion [10][11][12] to protein folding 13 , unfolding 14 , and selfassembly 15 .…”

Section: Introductionmentioning

confidence: 99%

Improving Molecular Force Fields Across Configurational Space by Combining Supervised and Unsupervised Machine Learning

Fonseca,

Poltavsky,

Vassilev-Galindo

et al. 2021

Preprint

View full text Add to dashboard Cite

The training set of atomic configurations is key to the performance of any Machine Learning Force Field (MLFF) and, as such, the training set selection determines the applicability of the MLFF model for predictive molecular simulations. However, most atomistic reference datasets are inhomogeneously distributed across configurational space (CS), thus choosing the training set randomly or according to the probability distribution of the data leads to models whose accuracy is mainly defined by the most common close-toequilibrium configurations in the reference data. In this work, we combine unsupervised and supervised ML methods to bypass the inherent bias of the data for common configurations, effectively widening the applicability range of MLFF to the fullest capabilities of the dataset. To achieve this goal, we first cluster the CS into subregions similar in terms of geometry and energetics. We iteratively test a given MLFF performance on each subregion and fill the training set of the model with the representatives of the most inaccurate parts of the CS. The proposed approach has been applied to a set of small organic molecules and alanine tetrapeptide, demonstrating an up to two-fold decrease in the root mean squared errors for force predictions of these molecules. This result holds for both kernel-based methods (sGDML and GAP/SOAP models) and deep neural networks (SchNet model).For the latter, the developed approach simultaneously improves both energy and forces, bypassing the compromise to be made when employing mixed energy/force loss functions.

show abstract