VQCrystal:利用矢量量化发现稳定晶体结构

ZiJie Qiu, Luozhijie Jin, Zijian Du, Hongyu Chen, Yan Cen, Siqi Sun, Yongfeng Mei, Hao Zhang
{"title":"VQCrystal:利用矢量量化发现稳定晶体结构","authors":"ZiJie Qiu, Luozhijie Jin, Zijian Du, Hongyu Chen, Yan Cen, Siqi Sun, Yongfeng Mei, Hao Zhang","doi":"arxiv-2409.06191","DOIUrl":null,"url":null,"abstract":"Discovering functional crystalline materials through computational methods\nremains a formidable challenge in materials science. Here, we introduce\nVQCrystal, an innovative deep learning framework that leverages discrete latent\nrepresentations to overcome key limitations in current approaches to crystal\ngeneration and inverse design. VQCrystal employs a hierarchical VQ-VAE\narchitecture to encode global and atom-level crystal features, coupled with a\nmachine learning-based inter-atomic potential(IAP) model and a genetic\nalgorithm to realize property-targeted inverse design. Benchmark evaluations on\ndiverse datasets demonstrate VQCrystal's advanced capabilities in\nrepresentation learning and novel crystal discovery. Notably, VQCrystal\nachieves state-of-the-art performance with 91.93\\% force validity and a\nFr\\'echet Distance of 0.152 on MP-20, indicating both strong validity and high\ndiversity in the sampling process. To demonstrate real-world applicability, we\napply VQCrystal for both 3D and 2D material design. For 3D materials, the\ndensity-functional theory validation confirmed that 63.04\\% of bandgaps and\n99\\% of formation energies of the 56 filtered materials matched the target\nrange. Moreover, 437 generated materials were validated as existing entries in\nthe full database outside the training set. For the discovery of 2D materials,\n73.91\\% of 23 filtered structures exhibited high stability with formation\nenergies below -1 eV/atom. Our results highlight VQCrystal's potential to\naccelerate the discovery of novel materials with tailored properties.","PeriodicalId":501234,"journal":{"name":"arXiv - PHYS - Materials Science","volume":null,"pages":null},"PeriodicalIF":0.0000,"publicationDate":"2024-09-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"VQCrystal: Leveraging Vector Quantization for Discovery of Stable Crystal Structures\",\"authors\":\"ZiJie Qiu, Luozhijie Jin, Zijian Du, Hongyu Chen, Yan Cen, Siqi Sun, Yongfeng Mei, Hao Zhang\",\"doi\":\"arxiv-2409.06191\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"Discovering functional crystalline materials through computational methods\\nremains a formidable challenge in materials science. Here, we introduce\\nVQCrystal, an innovative deep learning framework that leverages discrete latent\\nrepresentations to overcome key limitations in current approaches to crystal\\ngeneration and inverse design. VQCrystal employs a hierarchical VQ-VAE\\narchitecture to encode global and atom-level crystal features, coupled with a\\nmachine learning-based inter-atomic potential(IAP) model and a genetic\\nalgorithm to realize property-targeted inverse design. Benchmark evaluations on\\ndiverse datasets demonstrate VQCrystal's advanced capabilities in\\nrepresentation learning and novel crystal discovery. Notably, VQCrystal\\nachieves state-of-the-art performance with 91.93\\\\% force validity and a\\nFr\\\\'echet Distance of 0.152 on MP-20, indicating both strong validity and high\\ndiversity in the sampling process. To demonstrate real-world applicability, we\\napply VQCrystal for both 3D and 2D material design. For 3D materials, the\\ndensity-functional theory validation confirmed that 63.04\\\\% of bandgaps and\\n99\\\\% of formation energies of the 56 filtered materials matched the target\\nrange. Moreover, 437 generated materials were validated as existing entries in\\nthe full database outside the training set. For the discovery of 2D materials,\\n73.91\\\\% of 23 filtered structures exhibited high stability with formation\\nenergies below -1 eV/atom. Our results highlight VQCrystal's potential to\\naccelerate the discovery of novel materials with tailored properties.\",\"PeriodicalId\":501234,\"journal\":{\"name\":\"arXiv - PHYS - Materials Science\",\"volume\":null,\"pages\":null},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2024-09-10\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"arXiv - PHYS - Materials Science\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/arxiv-2409.06191\",\"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 - Materials Science","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/arxiv-2409.06191","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 0

摘要

通过计算方法发现功能晶体材料仍然是材料科学领域的一项艰巨挑战。在这里,我们介绍一种创新的深度学习框架--VQCrystal,它利用离散的潜在表征来克服当前晶体生成和逆向设计方法中的关键局限。VQCrystal 采用分层 VQ-VAE 架构来编码全局和原子级别的晶体特征,并结合基于机器学习的原子间势(IAP)模型和基因算法来实现以属性为目标的逆向设计。在各种数据集上进行的基准评估证明了 VQCrystal 在表征学习和新型晶体发现方面的先进能力。值得注意的是,VQCrystal在MP-20上达到了91.93%的受力有效率和0.152的Fr/'echet Distance,表明其在采样过程中具有很高的有效性和多样性。为了证明其在现实世界中的适用性,我们将 VQCrystal 应用于三维和二维材料设计。对于三维材料,密度-函数理论验证证实 56 种过滤材料中 63.04% 的带隙和 99% 的形成能符合目标范围。此外,437 种生成的材料被验证为训练集之外的完整数据库中的现有条目。在发现二维材料方面,23 种过滤结构中有 73.91% 的结构表现出高稳定性,其形成能低于-1 eV/原子。我们的研究结果凸显了 VQCrystal 在加速发现具有定制特性的新型材料方面的潜力。
本文章由计算机程序翻译,如有差异,请以英文原文为准。
VQCrystal: Leveraging Vector Quantization for Discovery of Stable Crystal Structures
Discovering functional crystalline materials through computational methods remains a formidable challenge in materials science. Here, we introduce VQCrystal, an innovative deep learning framework that leverages discrete latent representations to overcome key limitations in current approaches to crystal generation and inverse design. VQCrystal employs a hierarchical VQ-VAE architecture to encode global and atom-level crystal features, coupled with a machine learning-based inter-atomic potential(IAP) model and a genetic algorithm to realize property-targeted inverse design. Benchmark evaluations on diverse datasets demonstrate VQCrystal's advanced capabilities in representation learning and novel crystal discovery. Notably, VQCrystal achieves state-of-the-art performance with 91.93\% force validity and a Fr\'echet Distance of 0.152 on MP-20, indicating both strong validity and high diversity in the sampling process. To demonstrate real-world applicability, we apply VQCrystal for both 3D and 2D material design. For 3D materials, the density-functional theory validation confirmed that 63.04\% of bandgaps and 99\% of formation energies of the 56 filtered materials matched the target range. Moreover, 437 generated materials were validated as existing entries in the full database outside the training set. For the discovery of 2D materials, 73.91\% of 23 filtered structures exhibited high stability with formation energies below -1 eV/atom. Our results highlight VQCrystal's potential to accelerate the discovery of novel materials with tailored properties.
求助全文
通过发布文献求助,成功后即可免费获取论文全文。 去求助
来源期刊
自引率
0.00%
发文量
0
×
引用
GB/T 7714-2015
复制
MLA
复制
APA
复制
导出至
BibTeX EndNote RefMan NoteFirst NoteExpress
×
提示
您的信息不完整,为了账户安全,请先补充。
现在去补充
×
提示
您因"违规操作"
具体请查看互助需知
我知道了
×
提示
确定
请完成安全验证×
copy
已复制链接
快去分享给好友吧!
我知道了
右上角分享
点击右上角分享
0
联系我们:info@booksci.cn Book学术提供免费学术资源搜索服务,方便国内外学者检索中英文文献。致力于提供最便捷和优质的服务体验。 Copyright © 2023 布克学术 All rights reserved.
京ICP备2023020795号-1
ghs 京公网安备 11010802042870号
Book学术文献互助
Book学术文献互助群
群 号:481959085
Book学术官方微信