Stefaan S. P. Hessmann, Kristof T. Schütt, Niklas W. A. Gebauer, Michael Gastegger, Tamio Oguchi, Tomoki Yamashita
{"title":"通过神经网络主动学习加速晶体结构搜索,实现快速松弛","authors":"Stefaan S. P. Hessmann, Kristof T. Schütt, Niklas W. A. Gebauer, Michael Gastegger, Tamio Oguchi, Tomoki Yamashita","doi":"arxiv-2408.04073","DOIUrl":null,"url":null,"abstract":"Global optimization of crystal compositions is a significant yet\ncomputationally intensive method to identify stable structures within chemical\nspace. The specific physical properties linked to a three-dimensional atomic\narrangement make this an essential task in the development of new materials. We\npresent a method that efficiently uses active learning of neural network force\nfields for structure relaxation, minimizing the required number of steps in the\nprocess. This is achieved by neural network force fields equipped with\nuncertainty estimation, which iteratively guide a pool of randomly generated\ncandidates towards their respective local minima. Using this approach, we are\nable to effectively identify the most promising candidates for further\nevaluation using density functional theory (DFT). Our method not only reliably\nreduces computational costs by up to two orders of magnitude across the\nbenchmark systems Si16 , Na8Cl8 , Ga8As8 and Al4O6 , but also excels in finding\nthe most stable minimum for the unseen, more complex systems Si46 and Al16O24 .\nMoreover, we demonstrate at the example of Si16 that our method can find\nmultiple relevant local minima while only adding minor computational effort.","PeriodicalId":501369,"journal":{"name":"arXiv - PHYS - Computational Physics","volume":"58 1","pages":""},"PeriodicalIF":0.0000,"publicationDate":"2024-08-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"Accelerating crystal structure search through active learning with neural networks for rapid relaxations\",\"authors\":\"Stefaan S. P. Hessmann, Kristof T. Schütt, Niklas W. A. Gebauer, Michael Gastegger, Tamio Oguchi, Tomoki Yamashita\",\"doi\":\"arxiv-2408.04073\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"Global optimization of crystal compositions is a significant yet\\ncomputationally intensive method to identify stable structures within chemical\\nspace. The specific physical properties linked to a three-dimensional atomic\\narrangement make this an essential task in the development of new materials. We\\npresent a method that efficiently uses active learning of neural network force\\nfields for structure relaxation, minimizing the required number of steps in the\\nprocess. This is achieved by neural network force fields equipped with\\nuncertainty estimation, which iteratively guide a pool of randomly generated\\ncandidates towards their respective local minima. Using this approach, we are\\nable to effectively identify the most promising candidates for further\\nevaluation using density functional theory (DFT). Our method not only reliably\\nreduces computational costs by up to two orders of magnitude across the\\nbenchmark systems Si16 , Na8Cl8 , Ga8As8 and Al4O6 , but also excels in finding\\nthe most stable minimum for the unseen, more complex systems Si46 and Al16O24 .\\nMoreover, we demonstrate at the example of Si16 that our method can find\\nmultiple relevant local minima while only adding minor computational effort.\",\"PeriodicalId\":501369,\"journal\":{\"name\":\"arXiv - PHYS - Computational Physics\",\"volume\":\"58 1\",\"pages\":\"\"},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2024-08-07\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"arXiv - PHYS - Computational Physics\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/arxiv-2408.04073\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"arXiv - PHYS - Computational Physics","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/arxiv-2408.04073","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
Accelerating crystal structure search through active learning with neural networks for rapid relaxations
Global optimization of crystal compositions is a significant yet
computationally intensive method to identify stable structures within chemical
space. The specific physical properties linked to a three-dimensional atomic
arrangement make this an essential task in the development of new materials. We
present a method that efficiently uses active learning of neural network force
fields for structure relaxation, minimizing the required number of steps in the
process. This is achieved by neural network force fields equipped with
uncertainty estimation, which iteratively guide a pool of randomly generated
candidates towards their respective local minima. Using this approach, we are
able to effectively identify the most promising candidates for further
evaluation using density functional theory (DFT). Our method not only reliably
reduces computational costs by up to two orders of magnitude across the
benchmark systems Si16 , Na8Cl8 , Ga8As8 and Al4O6 , but also excels in finding
the most stable minimum for the unseen, more complex systems Si46 and Al16O24 .
Moreover, we demonstrate at the example of Si16 that our method can find
multiple relevant local minima while only adding minor computational effort.