Skip to main content

Thank you for visiting You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

Improving the efficiency of ab initio electronic-structure calculations by deep learning

A deep neural network method is developed to learn the density functional theory (DFT) Hamiltonian as a function of atomic structure. This approach provides a solution to the accuracy–efficiency dilemma of DFT and opens opportunities to investigate large-scale materials, such as twisted van der Waals materials.

Your institute does not have access to this article

Access options

Buy article

Get time limited or full article access on ReadCube.


All prices are NET prices.

Fig. 1: The DeepH method for efficient ab initio electronic-structure calculations.


  1. Jones, R. O. Density functional theory: Its origins, rise to prominence, and future. Rev. Mod. Phys. 87, 897 (2015). A review article that presents the development and important applications of DFT.

    MathSciNet  Article  Google Scholar 

  2. Butler, K. T. et al. Machine learning for molecular and materials science. Nature 559, 547–555 (2018). A review article that presents recent progress in machine learning for computational materials science and chemistry.

    Article  Google Scholar 

  3. Unke, O. T. et al. Machine learning force fields. Chem. Rev. 121, 10142–10186 (2021). A review article that presents applications of machine learning methods in the prediction of total energy to perform efficient molecular dynamics simulations.

    Article  Google Scholar 

  4. Carr, S. et al. Electronic-structure methods for twisted moiré layers. Nat. Rev. Mater. 5, 748–763 (2020). A review article that summarizes and compares different theoretical methods used to study twisted materials.

    Article  Google Scholar 

  5. Zhang, L. et al. Deep potential molecular dynamics: A scalable model with the accuracy of quantum mechanics. Phys. Rev. Lett. 120, 143001 (2018). An article that introduces a deep-learning approach to predicting the many-body potential.

    Article  Google Scholar 

Download references

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

This is a summary of: Li, H. et al. Deep-learning density functional theory Hamiltonian for efficient ab initio electronic-structure calculation. Nat. Comput. Sci. (2022).

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Improving the efficiency of ab initio electronic-structure calculations by deep learning. Nat Comput Sci 2, 418–419 (2022).

Download citation

  • Published:

  • Issue Date:

  • DOI:


Quick links

Nature Briefing

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing