Titel
Optimizing the architecture of Behler–Parrinello neural network potentials
Abstract
The architecture of neural network potentials is typically optimized at the beginning of the training process and remains unchanged throughout. Here, we investigate the accuracy of Behler–Parrinello neural network potentials for varying training set sizes. Using the QM9 and 3BPA datasets, we show that adjusting the network architecture according to the training set size improves the accuracy significantly. We demonstrate that both an insufficient and an excessive number of fitting parameters can have a detrimental impact on the accuracy of the neural network potential. Furthermore, we investigate the influences of descriptor complexity, neural network depth, and activation function on the model’s performance. We find that for the neural network potentials studied here, two hidden layers yield the best accuracy and that unbounded activation functions outperform bounded ones.
Stichwort
Atomistic simulationsPotential energy surfacesArtificial neural networksMachine learningSymmetry functions
Objekt-Typ
Sprache
Englisch [eng]
Persistent identifier
Erschienen in
Titel
The Journal of Chemical Physics
Band
159
Ausgabe
9
ISSN
0021-9606
Erscheinungsdatum
2023
Publication
AIP Publishing
Projekt
Kod / Identifikator
DOC 85
Erscheinungsdatum
2023
Zugänglichkeit
Rechte
Rechteangabe
© 2023 Author(s)
Universität Wien | Universitätsring 1 | 1010 Wien | T
+43-1-4277-0