Application of artificial neural networks for modeling of electronic excitation dynamics in 2D lattice: Direct and inverse problems
Source abstract
Abstract
Machine learning (ML) approaches are attracting wide interest in the chemical physics community since a trained ML system can predict numerical properties of various molecular systems with a small computational cost. In this work, we analyze the applicability of deep, sequential, and fully connected neural networks (NNs) to predict the excitation decay kinetics of a simple two-dimensional lattice model, which can be adapted to describe numerous real-life systems, such as aggregates of photosynthetic molecular complexes. After choosing a suitable loss function for NN training, we have achieved excellent accuracy for a direct problem—predictions of lattice excitation decay kinetics from the model parameter values. For an inverse problem—prediction of the model parameter values from the kinetics—we found that even though the kinetics obtained from estimated values differ from the actual ones, the values themselves are predicted with a reasonable accuracy. Finally, we discuss possibilities for applications of NNs for solving global optimization problems that are related to the need to fit experimental data using similar models.
Source: crossref · Quality: baseline
Research context
OpenAlex topics: Machine Learning in Materials Science · Spectroscopy and Quantum Chemical Studies · Advanced Chemical Physics Studies
SciVX discovery tags are terminology-based. OpenAlex topics are source-provided classifications.
Affiliations
Institute of Chemical Physics, Faculty of Physics, Vilnius University 1 , Saulėtekio 9-III, 10222 Vilnius, Lithuania
Department of Molecular Compound Physics, Center for Physical Sciences and Technology 2 , Saulėtekio 3, 10257 Vilnius, Lithuania
Funding
Lietuvos Mokslo Taryba
References (29)
Show source-provided references
1. Rev. Mod. Phys. · 2019 · 91 · 04500210.1103/revmodphys.91.045002 ↗
2. J. Phys. Chem. Lett. · 2020 · 11 · 233610.1021/acs.jpclett.9b03664 ↗
3. J. Phys. Chem. A · 2020 · 124 · 674510.1021/acs.jpca.0c04473 ↗
4. J. Phys. Chem. C · 2020 · 124 · 704810.1021/acs.jpcc.0c00329 ↗
5. J. Chem. Phys. · 2020 · 153 · 07411110.1063/5.0016009 ↗
6. Chem. Sci. · 2017 · 8 · 841910.1039/c7sc03542j ↗
7. Chem. Phys. · 2018 · 515 · 27210.1016/j.chemphys.2018.05.019 ↗
8. J. Phys. Chem. Lett. · 2021 · 12 · 247610.1021/acs.jpclett.1c00079 ↗
9. Chem. Phys. · 2019 · 520 · 5210.1016/j.chemphys.2019.01.002 ↗
10. J. Phys. Chem. Lett. · 2020 · 11 · 755910.1021/acs.jpclett.0c02168 ↗
11. 2016
12. 2008
13. 2010
14. 2005
15. Nat. Plants · 2016 · 2 · 1604510.1038/nplants.2016.45 ↗
16. Photosynth. Res. · 2018 · 135 · 27510.1007/s11120-017-0430-7 ↗
17. Chem. Sci. · 2020 · 11 · 569710.1039/d0sc00781a ↗
18. Adv. Mater. · 2009 · 21 · 274110.1002/adma.200803541 ↗
19. Phys. Chem. Chem. Phys. · 2022 · 24 · 1394110.1039/d2cp00557c ↗
20. J. Chem. Phys. · 1987 · 86 · 50310.1063/1.452301 ↗
21. J. Phys. Chem. C · 2012 · 116 · 374710.1021/jp209293k ↗
22. Biophys. J. · 2011 · 100 · 209410.1016/j.bpj.2011.03.049 ↗
23. 2014
24. D. P. Kingma and J. Ba, “Adam: A method for stochastic optimization,” arXiv:1412.6980 (2014).
25. Wallach · 2019 · 8024
26. Biochim. Biophys. Acta · 2019 · 1860 · 27110.1016/j.bbabio.2018.12.006 ↗
27. J. Phys. Chem. C · 2021 · 125 · 1309610.1021/acs.jpcc.1c02693 ↗
28. Chem. Rev. · 2022 · 122 · 425710.1021/acs.chemrev.1c00623 ↗
29. Phys. Rev. Res. · 2022 · 4 · 03317510.1103/physrevresearch.4.033175 ↗