Composition Based Oxidation State Prediction of Materials Using Deep Learning Language Models
Source abstract
Abstract
Abstract Oxidation states (OS) are the charges on atoms due to electrons gained or lost upon applying an ionic approximation to their bonds. As a fundamental property, OS has been widely used in charge‐neutrality verification, crystal structure determination, and reaction estimation. Currently, only heuristic rules exist for guessing the oxidation states of a given compound with many exceptions. Recent work has developed machine learning models based on heuristic structural features for predicting the oxidation states of metal ions. However, composition‐based oxidation state prediction still remains elusive so far, which has significant implications for the discovery of new materials for which the structures have not been determined. This work proposes a novel deep learning‐based BERT transformer language model BERTOS for predicting the oxidation states for all elements of inorganic compounds given only their chemical composition. This model achieves 96.82% accuracy for all‐element oxidation states prediction benchmarked on the cleaned ICSD dataset and achieves 97.61% accuracy for oxide materials. It is also demonstrated how it can be used to conduct large‐scale screening of hypothetical material compositions for materials discovery.
Source: crossref · Quality: baseline
Research context
OpenAlex topics: Machine Learning in Materials Science · X-ray Diffraction in Crystallography · Computational Drug Discovery Methods
SciVX discovery tags are terminology-based. OpenAlex topics are source-provided classifications.
Affiliations
Department of Computer Science and Engineering University of South Carolina Columbia SC 29201 USA
Dutch Fork High School Irmo SC 29063 USA
Hangzhou University of Electronic Science and Technology Hangzhou 311305 China
Department of Chemistry and Biochemistry University of South Carolina Columbia SC 29201 USA
Funding
National Science Foundation · Division of Materials Research
References (31)
Show source-provided references
1. 10.1038/s41563-018-0165-710.1038/s41563-018-0165-7 ↗
2. Jorgensen C. K. · 2012
3. 10.1038/s41524-020-00352-010.1038/s41524-020-00352-0 ↗
4. 10.1002/anie.20140756110.1002/anie.201407561 ↗
5. 10.1039/C8FD00032H10.1039/C8FD00032H ↗
6. L.Wei Q.Li Y.Song S.Stefanov E.Siriwardane F.Chen J.Hu arxiv:2204.11953 2022.
7. Fu N. · Mach. Learn. · 2023 · 4
8. 10.1038/s41557-021-00717-y10.1038/s41557-021-00717-y ↗
9. Amin M. · Photosynth. Res. · 2022
10. 10.1021/acs.jcim.2c0008010.1021/acs.jcim.2c00080 ↗
11. 10.1107/S205252061901304010.1107/S2052520619013040 ↗
12. 10.1016/j.commatsci.2019.01.04610.1016/j.commatsci.2019.01.046 ↗
13. 10.1016/j.commatsci.2012.10.02810.1016/j.commatsci.2012.10.028 ↗
14. J.Devlin M.‐W.Chang K.Lee K.Toutanova arXiv:1810.04805 2018.
15. 10.1038/s42256-022-00499-z10.1038/s42256-022-00499-z ↗
16. 10.1038/s41586-021-03819-210.1038/s41586-021-03819-2 ↗
17. 10.1093/bioinformatics/btac02010.1093/bioinformatics/btac020 ↗
18. Lin Z. · Science · 2023 · 379 · 6637
19. Vaswani A. · Adv. Neural Inf. Process. Syst. · 2017 · 30 · 5998
20. L.Wei N.Fu Y.Song Q.Wang J.Hu arXiv:2209.09406 2022.
21. 10.1107/S160057671900997X10.1107/S160057671900997X ↗
22. Deshmukh R. D. · 2020 · 76
23. Wolf T. · 2020 · 3810.18653/v1/2020.emnlp-demos.6 ↗
24. Muthuraman K. · 2021 · 5910.18653/v1/2021.dash-1.10 ↗
25. 10.1038/s41467-020-19964-710.1038/s41467-020-19964-7 ↗
26. 10.1007/s11837-013-0755-410.1007/s11837-013-0755-4 ↗
27. 10.1063/1.481232310.1063/1.4812323 ↗
28. 10.1080/00031305.1998.1048055910.1080/00031305.1998.10480559 ↗
29. Wei L. · Inorg. Chem · 2022 · 61 · 22
30. 10.1016/j.commatsci.2022.11149610.1016/j.commatsci.2022.111496 ↗
31. 10.1063/5.013352810.1063/5.0133528 ↗