[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-119048-en":3,"doc-seo-119048-105":30,"detail-sidebar-cat-0-en-105":91},{"code":4,"msg":5,"data":6},0,"success",{"doc_id":7,"user_id":8,"nickname":9,"user_avatar":10,"doc_module":4,"category_id":11,"category_name":12,"doc_title":13,"doc_description":14,"doc_content":15,"file_id":16,"file_url":17,"file_type":18,"file_size":19,"view_count":4,"is_deleted":4,"is_public":20,"is_downloadable":20,"audit_status":20,"page_count":21,"language":22,"language_code":23,"site_id":24,"html_lang":23,"table_of_contents":25,"faqs":26,"seo_title":27,"seo_description":14,"update_tm":28,"read_time":29},119048,137441390410,"Hazel","https://ap-avatar.wpscdn.com/avatar/2000252f4ab5702993?_k=1776741390130283984",8,"Research & Report","Transferable empirical pseudopotentials from machine learning","Machine learning is used to generate empirical pseudopotentials that represent local screened interactions in the Kohn-Sham Hamiltonian. The method introduces momentum-rangeseparated rotation-covariant descriptors to encode crystal symmetries and bond directionality, enabling accurate descriptions of anisotropic solids. Trained empirical potentials are shown to be versatile and transferable: energy bands and wave functions computed without self-consistency reproduce conventional ab initio results, including cases with defects in semiconductors. The framework supports integration into existing computational packages and extensions toward optical and transport properties.","arXiv :2306 .04426v2 [ cond-mat .mtrl-sci ] 7 Feb 2024  \nTransferable empirical pseudopotentials from machine learning  \nRokyeon Kim∗ and Young-Woo Son†  \nKorea Institute for Advanced Study, Seoul 02455, Korea  \n(Dated: February 8, 2024)  \nMachine learning is used to generate empirical pseudopotentials that characterize the local screened interactions in the Kohn-Sham Hamiltonian. Our approach incorporates momentum-rangeseparated rotation-covariant descriptors to capture crystal symmetries as well as crucial directional information of bonds, thus realizing accurate descriptions of anisotropic solids. Trained empirical potentials are shown to be versatile and transferable such that the calculated energy bands and wave functions without cumbersome self-consistency reproduce conventional ab initio results even for semiconductors with defects, thus fostering faster and faithful data-driven materials researches.  \nI. INTRODUCTION  \nFirst-principles calculations based on the density functional theory (DFT) [1, 2] have become standard tools for studying the physical properties of materials [3–6] . Recently, applications of machine learning (ML) techniques to various computational methodologies based on the DFT has brought forth a new set of tools for investigating materials at the quantum scale [7–11] . Such novel approaches have given rise to a rapidly growing field, offering new insights and significant potential for prediction and analysis of materials properties [12–27] .  \nOne of popular applications in those developments has been the accelerated computations of physical quantities such as total energies and atomic forces [12–14, 16– 21 , 27] . By circumventing a part of various computationally demanding processes involved in DFT calculations, the ML techniques provide efficient ways for improving various simulation methods such as molecular dynamics. On the other hand, the integration of ML has also brought improvements to the exchange-correlation functionals within DFT [15, 22–26, 28], which are central to describing the many-electron effects in quantum systems.  \nDespite these strides, applications of ML to obtain precise quantum mechanical electronic structures for entire phase space of interests remain relatively unexplored [29– 35] . Previous studies have utilized ML to study the electronic structures of one-dimensional [29], slab [30], and molecular [31–33] systems. We also note that the neural network was used to generate better transferable local pseudopotentials [36] . Only recently, a general ML framework to construct DFT Hamiltonian in the tightbinding approach has been developed [34, 35] . Accurate quantum properties of solids such as energy bands and wave functions are central to design and discovery of new materials with desired properties. However, the traditional DFT methods require a large amount of resources, partly because of the unavoidable self-consistent condition, posing significant challenges in data-intensive materials researches. Hence, there is a pressing need for  \n∗ Email: [rrykim@gmail.com](rrykim@gmail.com)  \n† Email: [hand@kias.re.kr](hand@kias.re.kr)  \na faster method that utilizes ML to accelerate electronic structure calculations without sacrificing the accuracy of first-principles methods.  \nBefore the advent of ab initio methods based on the DFT, empirical pseudopotential method (EPM) [37] has been widely used as a fast and efficient method to calculate the electronic structure of materials because of its formal simplicity as well as less demanding computational resources. Despite its extensive use for various solids [37–43], the EPM has limitations such as inaccurate wave functions [44] and transferability issues of the obtained pseudopotential [41, 42 , 45–47] . To improve EPM, Wang and Zunger proposed the localdensity-derived EPM, which generates pseudopotentials by inverting the Kohn-Sham (KS) potential in DFT calculations [48] . However, the potentials obtained from this approach still ","cbCaijEWmizaLdVn","https://ap.wps.com/l/cbCaijEWmizaLdVn","pdf",2733643,1,10,"English","en",105,"# Introduction\n## ML for accelerated DFT-related calculations\n## Limitations of existing empirical pseudopotential methods\n# Machine Learning Framework\n## Kohn-Sham equation and potential terms\n## Neural network generation of universal empirical pseudopotentials","[{\"question\":\"What is the main goal of the proposed approach?\",\"answer\":\"To use machine learning to construct transferable empirical pseudopotentials that capture local screened interactions in the Kohn-Sham Hamiltonian, enabling accurate electronic-structure predictions without self-consistency.\"},{\"question\":\"How does the method achieve better transferability to anisotropic and defective solids?\",\"answer\":\"It employs momentum-rangeseparated rotation-covariant descriptors that encode crystal symmetries and crucial bond direction information, supporting learned potentials that generalize to different target solids.\"},{\"question\":\"What computational advantage does the approach offer compared with conventional first-principles calculations?\",\"answer\":\"It reproduces energy bands and wave functions without the cumbersome self-consistent procedure required in standard DFT workflows, making data-driven materials research faster while retaining first-principles accuracy.\"}]","Transferable empirical pseudopotentials from machine learning | PDF",1785722083,25,{"code":4,"msg":31,"data":32},"ok",{"site_id":24,"language":23,"slug":33,"title":13,"keywords":34,"description":14,"schema_data":35,"social_meta":86,"head_meta":88,"extra_data":90,"updated_unix":28},"transferable-empirical-pseudopotentials-from-machine-learning","",{"@graph":36,"@context":85},[37,54,68],{"@type":38,"itemListElement":39},"BreadcrumbList",[40,44,48,51],{"item":41,"name":42,"@type":43,"position":20},"https://docshare.wps.com","Home","ListItem",{"item":45,"name":46,"@type":43,"position":47},"https://docshare.wps.com/document/","Document",2,{"item":49,"name":12,"@type":43,"position":50},"https://docshare.wps.com/document/research-report/",3,{"item":52,"name":13,"@type":43,"position":53},"https://docshare.wps.com/document/transferable-empirical-pseudopotentials-from-machine-learning/119048/",4,{"url":52,"name":13,"@type":55,"author":56,"headline":13,"publisher":58,"fileFormat":61,"inLanguage":23,"description":14,"dateModified":62,"datePublished":62,"encodingFormat":61,"isAccessibleForFree":63,"interactionStatistic":64},"DigitalDocument",{"name":9,"@type":57},"Person",{"url":41,"name":59,"@type":60},"DocShare","Organization","application/pdf","2026-08-03",true,{"@type":65,"interactionType":66,"userInteractionCount":4},"InteractionCounter",{"@type":67},"ViewAction",{"@type":69,"mainEntity":70},"FAQPage",[71,77,81],{"name":72,"@type":73,"acceptedAnswer":74},"What is the main goal of the proposed approach?","Question",{"text":75,"@type":76},"To use machine learning to construct transferable empirical pseudopotentials that capture local screened interactions in the Kohn-Sham Hamiltonian, enabling accurate electronic-structure predictions without self-consistency.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"How does the method achieve better transferability to anisotropic and defective solids?",{"text":80,"@type":76},"It employs momentum-rangeseparated rotation-covariant descriptors that encode crystal symmetries and crucial bond direction information, supporting learned potentials that generalize to different target solids.",{"name":82,"@type":73,"acceptedAnswer":83},"What computational advantage does the approach offer compared with conventional first-principles calculations?",{"text":84,"@type":76},"It reproduces energy bands and wave functions without the cumbersome self-consistent procedure required in standard DFT workflows, making data-driven materials research faster while retaining first-principles accuracy.","https://schema.org",{"og:url":52,"og:type":87,"og:title":13,"og:site_name":59,"og:description":14},"article",{"robots":89,"canonical":52},"index,follow",{"doc_id":7,"site_id":24},{"code":4,"msg":5,"data":92},[93,97,101,105,110,115,120,123,128,131,134],{"id":20,"doc_module":4,"doc_module_name":46,"category_name":94,"show_sort_weight":95,"slug":96},"Story & Novel",90,"story-novel",{"id":47,"doc_module":4,"doc_module_name":46,"category_name":98,"show_sort_weight":99,"slug":100},"Literature",80,"literature",{"id":53,"doc_module":4,"doc_module_name":46,"category_name":102,"show_sort_weight":103,"slug":104},"Exam",70,"exam",{"id":106,"doc_module":4,"doc_module_name":46,"category_name":107,"show_sort_weight":108,"slug":109},5,"Comic",60,"comic",{"id":111,"doc_module":4,"doc_module_name":46,"category_name":112,"show_sort_weight":113,"slug":114},6,"Technology",50,"technology",{"id":116,"doc_module":4,"doc_module_name":46,"category_name":117,"show_sort_weight":118,"slug":119},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":46,"category_name":12,"show_sort_weight":121,"slug":122},30,"research-report",{"id":124,"doc_module":4,"doc_module_name":46,"category_name":125,"show_sort_weight":126,"slug":127},9,"Religion & Spirituality",20,"religion-spirituality",{"id":126,"doc_module":4,"doc_module_name":46,"category_name":129,"show_sort_weight":126,"slug":130},"World Cup","world-cup",{"id":21,"doc_module":4,"doc_module_name":46,"category_name":132,"show_sort_weight":21,"slug":133},"Lifestyle","lifestyle",{"id":135,"doc_module":4,"doc_module_name":46,"category_name":136,"show_sort_weight":106,"slug":137},19,"General","general"]