[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-121093-en":3,"doc-seo-121093-105":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":20,"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},121093,8796095461564,"Liam","https://ap-avatar.wpscdn.com/davatar_155a257f0dc6eb9ab79c44ca47cae57d",8,"Research & Report","On-the-fly training of polynomial machine learning potentials in computing lattice thermal conductivity","First-principles calculations combined with the linearized phonon Boltzmann equation enable lattice thermal conductivity (LTC) prediction, but accurate results rely on determining force constants and remain computationally intensive. This work integrates polynomial machine learning potentials into an intermediate step of the LTC workflow, training on a small displacement–force and energy dataset and generating a larger dataset efficiently. A modular, well-optimized high-throughput workflow is implemented to compute LTCs for 103 wurtzite, zincblende, and rocksalt compounds, demonstrating substantial reductions in required computational resources.","arXiv :2401 . 17531v3 [ cond-mat .mtrl-sci ] 13 May 2024  \nOn-the-􀀍y training of polynomial machine learning potentials in computing lattice  \nthermal conductivity  \nAtsushi Togo 1, 􀀃 and Atsuto Seko2  \n1 Center for Basic Research on Materials National Institute for Materials Science, Tsukuba, Ibaraki 305-0047, Japan  \n2 Department of Materials Science and Engineering,  \nKyoto University, Sakyo, Kyoto 606-8501, Japan  \nThe application of 􀀌rst-principles calculations for predicting lattice thermal conductivity (LTC) in crystalline materials, in conjunction with the linearized phonon Boltzmann equation, has gained increasing popularity. In this calculation, the determination of force constants through 􀀌rst-principles calculations is critical for accurate LTC predictions. For material exploration, performing 􀀌rstprinciples LTC calculations in a high-throughput manner is now expected, although it requires signi􀀌cant computational resources. To reduce computational demands, we integrated polynomial machine learning potentials on-the-􀀍y during the 􀀌rst-principles LTC calculations. This paper presents a systematic approach to 􀀌rst-principles LTC calculations. We designed and optimized ane􀀎cient work􀀍ow that integrates multiple modular software packages. We applied this approach to calculate LTCs for 103 compounds of the wurtzite, zincblende, and rocksalt types to evaluate the performance of the polynomial machine learning potentials in LTC calculations. We demonstrate asigni􀀌cant reduction in the computational resources required for the LTC predictions.  \nI. INTRODUCTION  \nCalculations of lattice thermal conductivity (LTC) based on 􀀌rst-principles calculations and the linearizedphonon Boltzmann equation [1–4] have become increasingly popular in recent years. This is because su􀀎ciently accurate LTC values can be systematically predicted fora wide variety of crystals using available computer simulation packages.[5–10] These computational tools are expected to be applied in materials discovery within a high-throughput calculation environment. However, since 􀀌rst-principles LTC calculations are still computationally intensive, there is a need for the development of methodologies to reduce the computational demands. We conventionally employ a supercell approach combined with the 􀀌nite displacement method for 􀀌rstprinciples LTC calculations. Random or systematic displacements are introduced to the supercells, and the forces on atoms are calculated using 􀀌rst-principles calculations. Subsequently, supercell force constants are computed from the dataset composed of the displacementsand forces, and the LTC values are calculated from these supercell force constants. Many supercells with di􀀋erent displacement con􀀌gurations are often required to populate the tensor elements of the supercell force constants. The accuracy of predicting LTCs relies on the use of 􀀌rst-principles calculations to obtain the displacementforce dataset. However, this approach is computationally intensive. In order to achieve precise LTC predictions with lower computational resources, compressive sensing force constants calculation methods were developed, as reported in Refs. 11 and 12 . These methods employ regularized linear regression techniques to eliminate certain tensor elements of the supercell force constants,  \nthereby reducing the required size of the displacementforce dataset.  \nIn this study, we introduce another approach to reduce the computational demands of 􀀌rst-principles LTC calculations. We incorporate polynomial machine learning potentials (MLPs) [13, 14] into an intermediate stage of the LTC calculation process. The polynomial MLPs are trained using a small dataset of displacement-force pairs and energies derived from 􀀌rst-principles calculations. Subsequently, the polynomial MLPs generate a large displacement-force dataset to calculate supercell force constants with signi􀀌cantly lower computational demands than those required by 􀀌rst-principles calcul","cbCaioV3IugDV7ux","https://ap.wps.com/l/cbCaioV3IugDV7ux","pdf",482414,1,12,"English","en",105,"# Introduction\n## Force-constant dataset and computational cost\n## Polynomial machine learning potentials in the LTC workflow\n## High-throughput goals and software workflow overview","[{\"question\":\"Why are first-principles lattice thermal conductivity calculations computationally demanding?\",\"answer\":\"They require constructing force constants from displacement–force datasets, which typically needs many supercells and extensive first-principles calculations to accurately populate tensor elements.\"},{\"question\":\"How does on-the-fly training reduce the cost of LTC predictions?\",\"answer\":\"The method trains polynomial machine learning potentials using a small set of displacement–force pairs and energies, then uses the trained model to generate a larger displacement–force dataset for computing supercell force constants.\"},{\"question\":\"What materials and crystal types were used to evaluate performance?\",\"answer\":\"LTCs were computed for 103 compounds across wurtzite, zincblende, and rocksalt crystal structure types to assess the effectiveness of the polynomial MLPs.\"}]","On-the-fly training of polynomial machine learning potentials in computing lattice thermal conductivity | PDF",1785733685,30,{"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},"on-the-fly-training-of-polynomial-machine-learning-potentials-in-computing-lattice-thermal-conductivity","",{"@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/on-the-fly-training-of-polynomial-machine-learning-potentials-in-computing-lattice-thermal-conductivity/121093/",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":20},"InteractionCounter",{"@type":67},"ViewAction",{"@type":69,"mainEntity":70},"FAQPage",[71,77,81],{"name":72,"@type":73,"acceptedAnswer":74},"Why are first-principles lattice thermal conductivity calculations computationally demanding?","Question",{"text":75,"@type":76},"They require constructing force constants from displacement–force datasets, which typically needs many supercells and extensive first-principles calculations to accurately populate tensor elements.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"How does on-the-fly training reduce the cost of LTC predictions?",{"text":80,"@type":76},"The method trains polynomial machine learning potentials using a small set of displacement–force pairs and energies, then uses the trained model to generate a larger displacement–force dataset for computing supercell force constants.",{"name":82,"@type":73,"acceptedAnswer":83},"What materials and crystal types were used to evaluate performance?",{"text":84,"@type":76},"LTCs were computed for 103 compounds across wurtzite, zincblende, and rocksalt crystal structure types to assess the effectiveness of the polynomial MLPs.","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},[93,97,101,105,110,115,120,122,127,130,134],{"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":29,"slug":121},"research-report",{"id":123,"doc_module":4,"doc_module_name":46,"category_name":124,"show_sort_weight":125,"slug":126},9,"Religion & Spirituality",20,"religion-spirituality",{"id":125,"doc_module":4,"doc_module_name":46,"category_name":128,"show_sort_weight":125,"slug":129},"World Cup","world-cup",{"id":131,"doc_module":4,"doc_module_name":46,"category_name":132,"show_sort_weight":131,"slug":133},10,"Lifestyle","lifestyle",{"id":135,"doc_module":4,"doc_module_name":46,"category_name":136,"show_sort_weight":106,"slug":137},19,"General","general"]