DeePKS + ABACUS as a Bridge between Expensive Quantum Mechanical Models and Machine Learning PotentialsClick to copy article linkArticle link copied!
- Wenfei Li
- Qi Ou
- Yixiao ChenYixiao ChenProgram in Applied and Computational Mathematics, Princeton University, Princeton, New Jersey08544, United StatesMore by Yixiao Chen
- Yu CaoYu CaoHEDPS, CAPT, College of Engineering and School of Physics, Peking University, Beijing100871, P. R. ChinaMore by Yu Cao
- Renxi LiuRenxi LiuHEDPS, CAPT, College of Engineering and School of Physics, Peking University, Beijing100871, P. R. ChinaMore by Renxi Liu
- Chunyi ZhangChunyi ZhangDepartment of Physics, Temple University, Philadelphia, Pennsylvania19122, United StatesMore by Chunyi Zhang
- Daye Zheng
- Chun CaiChun CaiAI for Science Institute, Beijing100080, P. R. ChinaDP Technology, Beijing100080, P. R. ChinaMore by Chun Cai
- Xifan WuXifan WuDepartment of Physics, Temple University, Philadelphia, Pennsylvania19122, United StatesMore by Xifan Wu
- Han Wang*Han Wang*Email: [email protected]Laboratory of Computational Physics, Institute of Applied Physics and Computational Mathematics, Huayuan Road 6, Beijing100088, P. R. ChinaMore by Han Wang
- Mohan Chen*Mohan Chen*Email: [email protected]HEDPS, CAPT, College of Engineering and School of Physics, Peking University, Beijing100871, P. R. ChinaMore by Mohan Chen
- Linfeng Zhang*Linfeng Zhang*Email: [email protected]AI for Science Institute, Beijing100080, P. R. ChinaDP Technology, Beijing100080, P. R. ChinaMore by Linfeng Zhang
Abstract

Recently, the development of machine learning (ML) potentials has made it possible to perform large-scale and long-time molecular simulations with the accuracy of quantum mechanical (QM) models. However, for different levels of QM methods, such as density functional theory (DFT) at the meta-GGA level and/or with exact exchange, quantum Monte Carlo, etc., generating a sufficient amount of data for training an ML potential has remained computationally challenging due to their high cost. In this work, we demonstrate that this issue can be largely alleviated with Deep Kohn–Sham (DeePKS), an ML-based DFT model. DeePKS employs a computationally efficient neural network-based functional model to construct a correction term added upon a cheap DFT model. Upon training, DeePKS offers closely matched energies and forces compared with high-level QM method, but the number of training data required is orders of magnitude less than that required for training a reliable ML potential. As such, DeePKS can serve as a bridge between expensive QM models and ML potentials: one can generate a decent amount of high-accuracy QM data to train a DeePKS model and then use the DeePKS model to label a much larger amount of configurations to train an ML potential. This scheme for periodic systems is implemented in a DFT package ABACUS, which is open source and ready for use in various applications.
Cited By
Smart citations by scite.ai include citation statements extracted from the full text of the citing article. The number of the statements may be higher than the number of citations provided by ACS Publications if one paper cites another multiple times or lower if scite has not yet processed some of the citing articles.
This article is cited by 8 publications.
- Pengchao Zhang, Xuefei Xu. Propensity of Water Self-Ions at Air(Oil)–Water Interfaces Revealed by Deep Potential Molecular Dynamics with Enhanced Sampling. Langmuir 2025, 41
(5)
, 3675-3683. https://doi.org/10.1021/acs.langmuir.4c05004
- Wai-Pan Ng, Zili Zhang, Jun Yang. Accurate Neural Network Fine-Tuning Approach for Transferable Ab Initio Energy Prediction across Varying Molecular and Crystalline Scales. Journal of Chemical Theory and Computation 2025, Article ASAP.
- Pengchao Zhang, Axel Tosello Gardini, Xuefei Xu, Michele Parrinello. Intramolecular and Water Mediated Tautomerism of Solvated Glycine. Journal of Chemical Information and Modeling 2024, 64
(9)
, 3599-3604. https://doi.org/10.1021/acs.jcim.4c00273
- Sophie Baker, Joshua Pagotto, Timothy T. Duignan, Alister J. Page. High-Throughput Aqueous Electrolyte Structure Prediction Using IonSolvR and Equivariant Graph Neural Network Potentials. The Journal of Physical Chemistry Letters 2023, 14
(42)
, 9508-9515. https://doi.org/10.1021/acs.jpclett.3c01783
- Qi Ou, Ping Tuo, Wenfei Li, Xiaoxu Wang, Yixiao Chen, Linfeng Zhang. DeePKS Model for Halide Perovskites with the Accuracy of a Hybrid Functional. The Journal of Physical Chemistry C 2023, 127
(37)
, 18755-18764. https://doi.org/10.1021/acs.jpcc.3c04703
- Kyle Bystrom, Boris Kozinsky. Nonlocal machine-learned exchange functional for molecules and solids. Physical Review B 2024, 110
(7)
https://doi.org/10.1103/PhysRevB.110.075130
- Jin Xiao, YiXiao Chen, LinFeng Zhang, Han Wang, Tong Zhu. A machine learning-based high-precision density functional method for drug-like molecules. Artificial Intelligence Chemistry 2024, 2
(1)
, 100037. https://doi.org/10.1016/j.aichem.2023.100037
- Xiao Xie, Mengteng Chen, Yawei Yu, Jia Li, Mingwen Zhao, Linyang Li. Structural phase transitions, mechanical properties, and electronic band structures of room-temperature ferromagnetic monolayers ScMP2 (M = Mn and Cr). Results in Physics 2023, 53 , 106988. https://doi.org/10.1016/j.rinp.2023.106988
Article Views are the COUNTER-compliant sum of full text article downloads since November 2008 (both PDF and HTML) across all institutions and individuals. These metrics are regularly updated to reflect usage leading up to the last few days.
Citations are the number of other articles citing this article, calculated by Crossref and updated daily. Find more information about Crossref citation counts.
The Altmetric Attention Score is a quantitative measure of the attention that a research article has received online. Clicking on the donut icon will load a page at altmetric.com with additional details about the score and the social media presence for the given article. Find more information on the Altmetric Attention Score and how the score is calculated.