Sie können Bookmarks mittels Listen verwalten, loggen Sie sich dafür bitte in Ihr SLUB Benutzerkonto ein.
Medientyp:
E-Artikel
Titel:
How robust are modern graph neural network potentials in long and hot molecular dynamics simulations?
Beteiligte:
Stocker, Sina;
Gasteiger, Johannes;
Becker, Florian;
Günnemann, Stephan;
Margraf, Johannes T
Erschienen:
IOP Publishing, 2022
Erschienen in:
Machine Learning: Science and Technology, 3 (2022) 4, Seite 045010
Sprache:
Nicht zu entscheiden
DOI:
10.1088/2632-2153/ac9955
ISSN:
2632-2153
Entstehung:
Anmerkungen:
Beschreibung:
Abstract Graph neural networks (GNNs) have emerged as a powerful machine learning approach for the prediction of molecular properties. In particular, recently proposed advanced GNN models promise quantum chemical accuracy at a fraction of the computational cost. While the capabilities of such advanced GNNs have been extensively demonstrated on benchmark datasets, there have been few applications in real atomistic simulations. Here, we therefore put the robustness of GNN interatomic potentials to the test, using the recently proposed GemNet architecture as a testbed. Models are trained on the QM7-x database of organic molecules and used to perform extensive molecular dynamics simulations. We find that low test set errors are not sufficient for obtaining stable dynamics and that severe pathologies sometimes only become apparent after hundreds of ps of dynamics. Nonetheless, highly stable and transferable GemNet potentials can be obtained with sufficiently large training sets.