Mitigating error cancellation in density functional approximations via machine learning correction
The Journal of Chemical Physics · 07.08.2025 · 163 (5)
DOI: 10.1063/5.0267783 ↗Kurz gefasst
Fachartikel in der Zeitschrift The Journal of Chemical Physics, online erschienen am 04.08.2025. Verfasst von Zipeng An und 7 weiteren Autor:innen. Beteiligte Einrichtung: Hefei National Research Center for Physical Sciences at the Microscale and Synergetic Innovation Center of Quantum Information and Quantum Physics, U…. Die Arbeit wurde bisher 4-mal zitiert (Crossref) und nennt 110 Literaturangaben. Gefördert von: National Natural Science Foundation of China, AI for Science Foundation of Fudan University, Strategic Priority Research Program of the Chinese Academy of Sciences.
Abstract (Originalsprache)
The integration of machine learning (ML) with density functional theory has emerged as a promising strategy to enhance the accuracy of density functional methods. While practical implementations of density functional approximations (DFAs) often exploit error cancellation between chemical species to achieve high accuracy in thermochemical and kinetic energy predictions, this approach is inherently system-dependent, which severely limits the transferability of DFAs. To address this challenge, we developed a novel ML-based correction to the widely used B3LYP functional, directly targeting its deviations from the exact exchange-correlation functional. By utilizing highly accurate absolute energies as exclusive reference data, our approach eliminates the reliance on error cancellation. To optimize the ML model, we attribute errors to real-space pointwise contributions and design a double-cycle protocol that incorporates self-consistent field calculations into the training workflow. Numerical tests demonstrate that the ML model, trained solely on absolute energies, improves the accuracy of calculated relative energies, demonstrating that robust DFAs can be constructed without resorting to error cancellation. Comprehensive benchmarks further show that our ML-corrected B3LYP functional significantly outperforms the original B3LYP across diverse thermochemical and kinetic energy calculations, offering a versatile and superior alternative for practical applications.
Autor:innen und Institutionen
- Hefei National Research Center for Physical Sciences at the Microscale and Synergetic Innovation Center of Quantum Information and Quantum Physics, University of Science and Technology of China, Hefei, Anhui 230026, Chi…
- Department of Chemistry, University of Basel, Klingelbergstrasse 80, CH-4056 Basel, Switzerland.
- Department of Chemistry, Fudan University, Shanghai 200438, China.
- Department of Chemistry, The University of Hong Kong, Pokfulam Road, Hong Kong, China.
ORCID-Profile: Zipeng An, JingChun Wang, Zhiyu Li, Yalun Zheng, GuanHua Chen, Xiao Zheng
Förderung
- National Natural Science Foundation of China
- AI for Science Foundation of Fudan University
- Strategic Priority Research Program of the Chinese Academy of Sciences
- Hong Kong Quantum AI Lab, AIR@InnoHK of the Hong Kong Government
Zitieren
APA
An, Z., Wang, J., Zhang, Y., Li, Z., Wu, J., Zheng, Y., Chen, G., & Zheng, X. (2025). Mitigating error cancellation in density functional approximations via machine learning correction. The Journal of Chemical Physics, 163(5). https://doi.org/10.1063/5.0267783
Vancouver
An Z, Wang J, Zhang Y, Li Z, Wu J, Zheng Y, et al. Mitigating error cancellation in density functional approximations via machine learning correction. The Journal of Chemical Physics. 2025;163(5). doi:10.1063/5.0267783
BibTeX
@article{An2025,
author = {An, Zipeng and Wang, JingChun and Zhang, Yapeng and Li, Zhiyu and Wu, Jiang and Zheng, Yalun and Chen, GuanHua and Zheng, Xiao},
title = {Mitigating error cancellation in density functional approximations via machine learning correction},
journal = {The Journal of Chemical Physics},
year = {2025},
volume = {163},
number = {5},
doi = {10.1063/5.0267783}
}
EUVIMED