{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,11,21]],"date-time":"2026-11-21T19:21:55Z","timestamp":1795288915520,"version":"4.10.3"},"reference-count":48,"publisher":"Springer Science and Business Media LLC","issue":"1","license":[{"start":{"date-parts":[[2023,5,18]],"date-time":"2023-05-18T00:00:00Z","timestamp":1684368000000},"content-version":"tdm","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0"},{"start":{"date-parts":[[2023,5,18]],"date-time":"2023-05-18T00:00:00Z","timestamp":1684368000000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0"}],"content-domain":{"domain":["link.springer.com"],"crossmark-restriction":false},"short-container-title":["Nat Commun"],"abstract":"<jats:title>Abstract<\/jats:title><jats:p>The combination of deep learning and ab initio calculation has shown great promise in revolutionizing future scientific research, but how to design neural network models incorporating a priori knowledge and symmetry requirements is a key challenging subject. Here we propose an E(3)-equivariant deep-learning framework to represent density functional theory (DFT) Hamiltonian as a function of material structure, which can naturally preserve the Euclidean symmetry even in the presence of spin\u2013orbit coupling. Our DeepH-E3 method enables efficient electronic structure calculation at ab initio accuracy by learning from DFT data of small-sized structures, making the routine study of large-scale supercells (&gt;10<jats:sup>4<\/jats:sup> atoms) feasible. The method can reach sub-meV prediction accuracy at high training efficiency, showing state-of-the-art performance in our experiments. The work is not only of general significance to deep-learning method development but also creates opportunities for materials research, such as building a Moir\u00e9-twisted material database.<\/jats:p>","DOI":"10.1038\/s41467-023-38468-8","type":"journal-article","created":{"date-parts":[[2023,5,19]],"date-time":"2023-05-19T05:03:57Z","timestamp":1684472637000},"update-policy":"https:\/\/doi.org\/10.1007\/springer_crossmark_policy","source":"Crossref","is-referenced-by-count":146,"title":["General framework for E(3)-equivariant neural network representation of density functional theory Hamiltonian"],"prefix":"10.1038","volume":"14","author":[{"ORCID":"https:\/\/orcid.org\/0000-0002-6459-4288","authenticated-orcid":false,"given":"Xiaoxun","family":"Gong","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-5967-5251","authenticated-orcid":false,"given":"He","family":"Li","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Nianlong","family":"Zou","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Runzhang","family":"Xu","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-9685-2547","authenticated-orcid":false,"given":"Wenhui","family":"Duan","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-4844-2460","authenticated-orcid":false,"given":"Yong","family":"Xu","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"297","published-online":{"date-parts":[[2023,5,18]]},"reference":[{"key":"38468_CR1","doi-asserted-by":"publisher","first-page":"146401","DOI":"10.1103\/PhysRevLett.98.146401","volume":"98","author":"J Behler","year":"2007","unstructured":"Behler, J. & Parrinello, M. Generalized neural-network representation of high-dimensional potential-energy surfaces. Phys. Rev. Lett. 98, 146401 (2007).","journal-title":"Phys. Rev. Lett."},{"key":"38468_CR2","doi-asserted-by":"publisher","first-page":"143001","DOI":"10.1103\/PhysRevLett.120.143001","volume":"120","author":"L Zhang","year":"2018","unstructured":"Zhang, L., Han, J., Wang, H., Car, R. & E, W. Deep potential molecular dynamics: a scalable model with the accuracy of quantum mechanics. Phys. Rev. Lett. 120, 143001 (2018).","journal-title":"Phys. Rev. Lett."},{"key":"38468_CR3","doi-asserted-by":"publisher","first-page":"3192","DOI":"10.1039\/C6SC05720A","volume":"8","author":"JS Smith","year":"2017","unstructured":"Smith, J. S., Isayev, O. & Roitberg, A. E. ANI-1: an extensible neural network potential with DFT accuracy at force field computational cost. Chem. Sci. 8, 3192 (2017).","journal-title":"Chem. Sci."},{"key":"38468_CR4","doi-asserted-by":"publisher","first-page":"241722","DOI":"10.1063\/1.5019779","volume":"148","author":"KT Sch\u00fctt","year":"2018","unstructured":"Sch\u00fctt, K. T., Sauceda, H. E., Kindermans, P.-J., Tkatchenko, A. & M\u00fcller, K.-R. SchNet\u2014a deep learning architecture for molecules and materials. J. Chem. Phys. 148, 241722 (2018).","journal-title":"J. Chem. Phys."},{"key":"38468_CR5","doi-asserted-by":"publisher","DOI":"10.1038\/s41467-021-27504-0","volume":"12","author":"OT Unke","year":"2021","unstructured":"Unke, O. T. et al. Spookynet: learning force fields with electronic degrees of freedom and nonlocal effects. Nat. Commun. 12, 7273 (2021).","journal-title":"Nat. Commun."},{"key":"38468_CR6","unstructured":"Gasteiger, J., Becker, F. & G\u00fcnnemann, S. Gemnet: universal directional graph neural networks for molecules. Adv. Neural Inf. Process. Syst. 34, 6790\u20136802 (2021)."},{"key":"38468_CR7","doi-asserted-by":"publisher","DOI":"10.1038\/s41467-022-29939-5","volume":"13","author":"S Batzner","year":"2022","unstructured":"Batzner, S. et al. E(3)-equivariant graph neural networks for data-efficient and accurate interatomic potentials. Nat. Commun. 13, 2453 (2022).","journal-title":"Nat. Commun."},{"key":"38468_CR8","unstructured":"Duvenaud, D. K. et al. Convolutional networks on graphs for learning molecular fingerprints. Adv. Neural Inf. Process. Syst. 28, 2224\u20132232 (2015)."},{"key":"38468_CR9","unstructured":"Gilmer, J., Schoenholz, S. S., Riley, P. F., Vinyals, O. & Dahl, G. E. Neural message passing for quantum chemistry. PMLR 70, 1263\u20131272 (2017)."},{"key":"38468_CR10","doi-asserted-by":"publisher","first-page":"22","DOI":"10.1038\/s41524-019-0162-7","volume":"5","author":"A Chandrasekaran","year":"2019","unstructured":"Chandrasekaran, A. et al. Solving the electronic structure problem with machine learning. NPJ Comput. Mater. 5, 22 (2019).","journal-title":"NPJ Comput. Mater."},{"key":"38468_CR11","doi-asserted-by":"publisher","DOI":"10.1038\/s41467-019-12875-2","volume":"10","author":"KT Sch\u00fctt","year":"2019","unstructured":"Sch\u00fctt, K. T., Gastegger, M., Tkatchenko, A., M\u00fcller, K.-R. & Maurer, R. J. Unifying machine learning and quantum chemistry with a deep neural network for molecular wavefunctions. Nat. Commun. 10, 5024 (2019).","journal-title":"Nat. Commun."},{"key":"38468_CR12","unstructured":"Gasteiger, J., Gro\u00df, J. & G\u00fcnnemann, S. Directional message passing for molecular graphs. In Proceedings of the Eighth International Conference on Learning Representations, https:\/\/openreview.net\/forum?id=B1eWbxStPH (ICLR, 2020)."},{"key":"38468_CR13","unstructured":"Anderson, B., Hy, T. S. & Kondor, R. Cormorant: covariant molecular neural networks. Adv. Neural Inf. Process. Syst. 32, 14537\u201314546 (2019)."},{"key":"38468_CR14","unstructured":"Sch\u00fctt, K., Unke, O. & Gastegger, M. Equivariant message passing for the prediction of tensorial properties and molecular spectra. PMLR 139, 9377\u20139388 (2021)."},{"key":"38468_CR15","doi-asserted-by":"publisher","first-page":"e2205221119","DOI":"10.1073\/pnas.2205221119","volume":"119","author":"Z Qiao","year":"2022","unstructured":"Qiao, Z. et al. Informing geometric deep learning with electronic interactions to accelerate quantum chemistry. Proc. Natl Acad. Sci. USA 119, e2205221119 (2022).","journal-title":"Proc. Natl Acad. Sci. USA"},{"key":"38468_CR16","unstructured":"J\u00f8rgensen, P. B., Jacobsen, K. W. & Schmidt, M. N. Neural message passing with edge updates for predicting properties of molecules and material. Preprint at https:\/\/arxiv.org\/abs\/1806.03146 (2018)."},{"key":"38468_CR17","doi-asserted-by":"publisher","first-page":"145301","DOI":"10.1103\/PhysRevLett.120.145301","volume":"120","author":"T Xie","year":"2018","unstructured":"Xie, T. & Grossman, J. C. Crystal graph convolutional neural networks for an accurate and interpretable prediction of material properties. Phys. Rev. Lett. 120, 145301 (2018).","journal-title":"Phys. Rev. Lett."},{"key":"38468_CR18","doi-asserted-by":"publisher","first-page":"3678","DOI":"10.1021\/acs.jctc.9b00181","volume":"15","author":"OT Unke","year":"2019","unstructured":"Unke, O. T. & Meuwly, M. Physnet: a neural network for predicting energies, forces, dipole moments, and partial charges. J. Chem. Theory Comput. 15, 3678 (2019).","journal-title":"J. Chem. Theory Comput."},{"key":"38468_CR19","unstructured":"Su, M., Yang, J.-H., Xiang, H.-J. & Gong, X.-G. Efficient prediction of density functional theory Hamiltonian with graph neural network. Preprint at https:\/\/arxiv.org\/abs\/2205.05475 (2022)."},{"key":"38468_CR20","unstructured":"Cohen, T. & Welling, M. Group equivariant convolutional networks. PMLR 48, 2990\u20132999 (2016)."},{"key":"38468_CR21","unstructured":"Thomas, N. et al. Tensor field networks: rotation-and translation-equivariant neural networks for 3D point clouds. Preprint at https:\/\/arxiv.org\/abs\/1802.08219 (2022)."},{"key":"38468_CR22","unstructured":"Weiler, M., Geiger, M., Welling, M., Boomsma, W. & Cohen, T. S. 3D steerable CNNs: learning rotationally equivariant features in volumetric data. Adv. Neural Inf. Process. Syst. 31, 10381\u201310392 (2018)."},{"key":"38468_CR23","unstructured":"Kondor, R., Lin, Z. & Trivedi, S. Clebsch\u2013Gordan nets: a fully Fourier space spherical convolutional neural network. Adv. Neural Inf. Process. Syst. 31 ,10117\u201310126 (2018)."},{"key":"38468_CR24","unstructured":"Geiger, M. & Smidt, T. e3nn: Euclidean neural networks. Preprint at https:\/\/arxiv.org\/abs\/2207.09453 (2022)."},{"key":"38468_CR25","doi-asserted-by":"publisher","first-page":"367","DOI":"10.1038\/s43588-022-00265-6","volume":"2","author":"H Li","year":"2022","unstructured":"Li, H. et al. Deep-learning density functional theory Hamiltonian for efficient ab initio electronic-structure calculation. Nat. Comput. Sci. 2, 367 (2022).","journal-title":"Nat. Comput. Sci."},{"key":"38468_CR26","unstructured":"Unke, O. T. et al. SE(3)-equivariant prediction of molecular wavefunctions and electronic densities. Adv. Neural Inf. Process. Syst. 34, 14434\u201314447 (2021)."},{"key":"38468_CR27","doi-asserted-by":"publisher","first-page":"014115","DOI":"10.1063\/5.0072784","volume":"156","author":"J Nigam","year":"2022","unstructured":"Nigam, J., Willatt, M. J. & Ceriotti, M. Equivariant representations for molecular Hamiltonians and N -center atomic-scale properties. J. Chem. Phys. 156, 014115 (2022).","journal-title":"J. Chem. Phys."},{"key":"38468_CR28","doi-asserted-by":"publisher","first-page":"158","DOI":"10.1038\/s41524-022-00843-2","volume":"8","author":"L Zhang","year":"2022","unstructured":"Zhang, L. et al. Equivariant analytical mapping of first principles Hamiltonians to accurate and transferable materials models. NPJ Comput. Mater. 8, 158 (2022).","journal-title":"NPJ Comput. Mater."},{"key":"38468_CR29","doi-asserted-by":"publisher","first-page":"11635","DOI":"10.1073\/pnas.0505436102","volume":"102","author":"E Prodan","year":"2005","unstructured":"Prodan, E. & Kohn, W. Nearsightedness of electronic matter. Proc. Natl Acad. Sci. USA 102, 11635 (2005).","journal-title":"Proc. Natl Acad. Sci. USA"},{"key":"38468_CR30","unstructured":"Battaglia, P. W. et al. Relational inductive biases, deep learning, and graph networks. Preprint at https:\/\/arxiv.org\/abs\/1806.01261 (2018)."},{"key":"38468_CR31","doi-asserted-by":"publisher","unstructured":"Geiger, M. et al. e3nn\/e3nn: 2021-08-27. Zenodo https:\/\/doi.org\/10.5281\/zenodo.5292912 (2021).","DOI":"10.5281\/zenodo.5292912"},{"key":"38468_CR32","doi-asserted-by":"publisher","first-page":"80","DOI":"10.1038\/nature26154","volume":"556","author":"Y Cao","year":"2018","unstructured":"Cao, Y. et al. Correlated insulator behaviour at half-filling in magic-angle graphene superlattices. Nature 556, 80 (2018).","journal-title":"Nature"},{"key":"38468_CR33","doi-asserted-by":"publisher","first-page":"43","DOI":"10.1038\/nature26160","volume":"556","author":"Y Cao","year":"2018","unstructured":"Cao, Y. et al. Unconventional superconductivity in magic-angle graphene superlattices. Nature 556, 43 (2018).","journal-title":"Nature"},{"key":"38468_CR34","doi-asserted-by":"publisher","first-page":"066401","DOI":"10.1103\/PhysRevLett.126.066401","volume":"126","author":"B Liu","year":"2021","unstructured":"Liu, B. et al. Higher-order band topology in twisted Moir\u00e9 superlattice. Phys. Rev. Lett. 126, 066401 (2021).","journal-title":"Phys. Rev. Lett."},{"key":"38468_CR35","doi-asserted-by":"publisher","first-page":"195419","DOI":"10.1103\/PhysRevB.99.195419","volume":"99","author":"P Lucignano","year":"2019","unstructured":"Lucignano, P., Alf\u00e8, D., Cataudella, V., Ninno, D. & Cantele, G. Crucial role of atomic corrugation on the flat bands and energy gaps of twisted bilayer graphene at the magic angle \u03b8\u2009~\u20091.08\u2218. Phys. Rev. B 99, 195419 (2019).","journal-title":"Phys. Rev. B"},{"key":"38468_CR36","doi-asserted-by":"publisher","first-page":"1674","DOI":"10.1143\/JPSJ.74.1674","volume":"74","author":"T Fukui","year":"2005","unstructured":"Fukui, T., Hatsugai, Y. & Suzuki, H. Chern numbers in discretized Brillouin zone: efficient method of computing (spin) hall conductances. J. Phys. Soc. Jpn. 74, 1674 (2005).","journal-title":"J. Phys. Soc. Jpn."},{"key":"38468_CR37","doi-asserted-by":"publisher","first-page":"165502","DOI":"10.1088\/0953-8984\/24\/16\/165502","volume":"24","author":"T Hoshi","year":"2012","unstructured":"Hoshi, T., Yamamoto, S., Fujiwara, T., Sogabe, T. & Zhang, S.-L. An order-N electronic structure theory with generalized eigenvalue equations and its application to a ten-million-atom system. J. Phys. Condens. Matter 24, 165502 (2012).","journal-title":"J. Phys. Condens. Matter"},{"key":"38468_CR38","doi-asserted-by":"publisher","first-page":"015003","DOI":"10.1103\/RevModPhys.89.015003","volume":"89","author":"F Giustino","year":"2017","unstructured":"Giustino, F. Electron-phonon interactions from first principles. Rev. Mod. Phys. 89, 015003 (2017).","journal-title":"Rev. Mod. Phys."},{"key":"38468_CR39","doi-asserted-by":"publisher","unstructured":"Gong, X. et al. Dataset1 for \u201cGeneral framework for E(3)-equivariant neural network representation of density functional theory Hamiltonian\u201d. Zenodo https:\/\/doi.org\/10.5281\/zenodo.7553640 (2023).","DOI":"10.5281\/zenodo.7553640"},{"key":"38468_CR40","doi-asserted-by":"publisher","unstructured":"Gong, X. et al. Dataset2 for \u201cGeneral framework for E(3)-equivariant neural network representation of density functional theory Hamiltonian\u201d. Zenodo https:\/\/doi.org\/10.5281\/zenodo.7553827 (2023).","DOI":"10.5281\/zenodo.7553827"},{"key":"38468_CR41","doi-asserted-by":"publisher","unstructured":"Gong, X. et al. Dataset3 for \u201cGeneral framework for E(3)-equivariant neural network representation of density functional theory Hamiltonian\u201d. Zenodo https:\/\/doi.org\/10.5281\/zenodo.7553843 (2023).","DOI":"10.5281\/zenodo.7553843"},{"key":"38468_CR42","doi-asserted-by":"publisher","first-page":"11169","DOI":"10.1103\/PhysRevB.54.11169","volume":"54","author":"G Kresse","year":"1996","unstructured":"Kresse, G. & Furthm\u00fcller, J. Efficient iterative schemes for ab initio total-energy calculations using a plane-wave basis set. Phys. Rev. B 54, 11169 (1996).","journal-title":"Phys. Rev. B"},{"key":"38468_CR43","doi-asserted-by":"publisher","first-page":"1396","DOI":"10.1103\/PhysRevLett.78.1396","volume":"78","author":"JP Perdew","year":"1997","unstructured":"Perdew, J. P., Burke, K. & Ernzerhof, M. Generalized gradient approximation made simple. Phys. Rev. Lett. 78, 1396 (1997).","journal-title":"Phys. Rev. Lett."},{"key":"38468_CR44","doi-asserted-by":"publisher","first-page":"17953","DOI":"10.1103\/PhysRevB.50.17953","volume":"50","author":"PE Bl\u00f6chl","year":"1994","unstructured":"Bl\u00f6chl, P. E. Projector augmented-wave method. Phys. Rev. B 50, 17953 (1994).","journal-title":"Phys. Rev. B"},{"key":"38468_CR45","doi-asserted-by":"publisher","first-page":"1758","DOI":"10.1103\/PhysRevB.59.1758","volume":"59","author":"G Kresse","year":"1999","unstructured":"Kresse, G. & Joubert, D. From ultrasoft pseudopotentials to the projector augmented-wave method. Phys. Rev. B 59, 1758 (1999).","journal-title":"Phys. Rev. B"},{"key":"38468_CR46","doi-asserted-by":"publisher","first-page":"154101","DOI":"10.1063\/1.2065267","volume":"123","author":"AD Becke","year":"2005","unstructured":"Becke, A. D. & Johnson, E. R. A density-functional model of the dispersion interaction. J. Chem. Phys. 123, 154101 (2005).","journal-title":"J. Chem. Phys."},{"key":"38468_CR47","unstructured":"Paszke, A. et al. Pytorch: an imperative style, high-performance deep learning library. Adv. Neural Inf. Process. Syst. 32, 8024\u20138035 (2019)."},{"key":"38468_CR48","doi-asserted-by":"publisher","unstructured":"Gong, X. et al. Code for \u201cGeneral framework for E(3)-equivariant neural network representation of density functional theory Hamiltonian\u201d. Zenodo https:\/\/doi.org\/10.5281\/zenodo.7554314 (2023).","DOI":"10.5281\/zenodo.7554314"}],"container-title":["Nature Communications"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/www.nature.com\/articles\/s41467-023-38468-8.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/www.nature.com\/articles\/s41467-023-38468-8","content-type":"text\/html","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/www.nature.com\/articles\/s41467-023-38468-8.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2023,5,19]],"date-time":"2023-05-19T05:04:30Z","timestamp":1684472670000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.nature.com\/articles\/s41467-023-38468-8"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2023,5,18]]},"references-count":48,"journal-issue":{"issue":"1","published-online":{"date-parts":[[2023,12]]}},"alternative-id":["38468"],"URL":"https:\/\/doi.org\/10.1038\/s41467-023-38468-8","relation":{},"ISSN":["2041-1723"],"issn-type":[{"value":"2041-1723","type":"electronic"}],"subject":[],"published":{"date-parts":[[2023,5,18]]},"assertion":[{"value":"1 November 2022","order":1,"name":"received","label":"Received","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"27 April 2023","order":2,"name":"accepted","label":"Accepted","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"18 May 2023","order":3,"name":"first_online","label":"First Online","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"The authors declare no competing interests.","order":1,"name":"Ethics","group":{"name":"EthicsHeading","label":"Competing interests"}}],"article-number":"2848"}}