Publications Repository - Helmholtz-Zentrum Dresden-Rossendorf
1 PublicationFinding Machine-Learning Surrogates for Electronic Structures without Training
Fiedler, L.; Hoffmann, N.; Mohammed, P.; Popoola, G. A.; Yovell, T.; Oles, V.; Ellis, J. A.; Rajamanickam, S.; Cangi, A.
A myriad of phenomena in materials science and chemistry rely on quantum-level simulations of the electronic structure in matter. While moving to larger length and time scales has been a pressing issue for decades, such large-scale electronic structure calculations are still challenging despite modern software approaches and advances in high-performance computing.
The silver lining in this regard is the use of machine learning to accelerate electronic structure calculations -- this line of research has recently gained growing attention.
The grand challenge therein is finding a suitable machine-learning model during a process called hyperparameter optimization. This, however, causes a massive computational overhead in addition to that of data generation.
We accelerate the construction of machine-learning surrogate models by roughly two orders of magnitude by circumventing excessive training during the hyperparameter optimization phase. We demonstrate our workflow for Kohn-Sham density functional theory, the most popular computational method in materials science and chemistry.
Keywords: Machine learning; Neural networks; Materials science
Related publications
-
Training-free hyperparameter optimization of neural networks for electronic …
ROBIS: 35225 is new version of this (Id 34368) publication -
Data publication: Training-free hyperparameter optimization of neural networks …
ROBIS: 35227 HZDR-primary research data are used by this (Id 34368) publication
-
Contribution to WWW
arXiv:2202.09186 [cond-mat.mtrl-sci]: https://arxiv.org/abs/2202.09186
Permalink: https://www.hzdr.de/publications/Publ-34368