An explainable “family bucket” model for simultaneous prediction of K-edge XANES for multiple light transition metals

IF 7.4 1区 化学 Q1 CHEMISTRY, MULTIDISCIPLINARY
Chenyu Huang, Yunjiang Zhang, Shuyuan Li, Huimin Wang, Yaxin Wang, Shihao Wei, Shaorui Sun
{"title":"An explainable “family bucket” model for simultaneous prediction of K-edge XANES for multiple light transition metals","authors":"Chenyu Huang, Yunjiang Zhang, Shuyuan Li, Huimin Wang, Yaxin Wang, Shihao Wei, Shaorui Sun","doi":"10.1039/d5sc00494b","DOIUrl":null,"url":null,"abstract":"X-ray near-edge structure (XANES) is a crucial bridge between the local structures and chemical properties of materials. Although there have been a number of studies devoted to the development of predictive K-edge XANES spectral models, existing methods are usually still limited to separate modeling for a specific absorbing element. Currently, there is a lack of a K-edge XANES spectra prediction model that can be broadly applied to a wide range of elements, which would enable data dispersed in terms of absorbing elements to be integrated and well utilized. In this work, we develop an innovative “family bucket” model based on a multi-head graph attention convolutional neural network by combining a multi-element mixed dataset with a crystal topology approach for the localized environment. The model is able to predict the K-edge XANES spectra for a wide range of light transition metals (periods 3 and 4) simultaneously. Moreover, it is demonstrated that the training scheme not only improves the accuracy of the model but also the efficiency of its training. In terms of interpretability, several fascinating insights were gained, uncovering the underlying mechanisms of the model for spectral prediction. We investigate the collective behavior of neurons by employing a range of responses to different samples as descriptive features. Notably, the analysis revealed that neurons in the neural network exhibit functional differentiation characteristics analogous to Brodmann areas in the cerebral cortex. The homology of data analysis indicates that the mutual learning of samples from different absorbing elements is occurring between close elements of the same period. Additionally, the attention scores of the samples are determined by both the absorbing element and its surrounding atomic environment. In conclusion, this research advances the understanding of the relationship between XANES spectra and material structures while providing valuable insights into neural networks, enhancing the comprehension of neuronal behavior.","PeriodicalId":9909,"journal":{"name":"Chemical Science","volume":"11 1","pages":""},"PeriodicalIF":7.4000,"publicationDate":"2025-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Chemical Science","FirstCategoryId":"92","ListUrlMain":"https://doi.org/10.1039/d5sc00494b","RegionNum":1,"RegionCategory":"化学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q1","JCRName":"CHEMISTRY, MULTIDISCIPLINARY","Score":null,"Total":0}
引用次数: 0

Abstract

X-ray near-edge structure (XANES) is a crucial bridge between the local structures and chemical properties of materials. Although there have been a number of studies devoted to the development of predictive K-edge XANES spectral models, existing methods are usually still limited to separate modeling for a specific absorbing element. Currently, there is a lack of a K-edge XANES spectra prediction model that can be broadly applied to a wide range of elements, which would enable data dispersed in terms of absorbing elements to be integrated and well utilized. In this work, we develop an innovative “family bucket” model based on a multi-head graph attention convolutional neural network by combining a multi-element mixed dataset with a crystal topology approach for the localized environment. The model is able to predict the K-edge XANES spectra for a wide range of light transition metals (periods 3 and 4) simultaneously. Moreover, it is demonstrated that the training scheme not only improves the accuracy of the model but also the efficiency of its training. In terms of interpretability, several fascinating insights were gained, uncovering the underlying mechanisms of the model for spectral prediction. We investigate the collective behavior of neurons by employing a range of responses to different samples as descriptive features. Notably, the analysis revealed that neurons in the neural network exhibit functional differentiation characteristics analogous to Brodmann areas in the cerebral cortex. The homology of data analysis indicates that the mutual learning of samples from different absorbing elements is occurring between close elements of the same period. Additionally, the attention scores of the samples are determined by both the absorbing element and its surrounding atomic environment. In conclusion, this research advances the understanding of the relationship between XANES spectra and material structures while providing valuable insights into neural networks, enhancing the comprehension of neuronal behavior.

Abstract Image

一个可解释的“家族桶”模型,用于同时预测多种轻过渡金属的k边XANES
x射线近边结构(XANES)是材料局部结构与化学性质之间的重要桥梁。虽然已经有许多研究致力于开发预测k边XANES光谱模型,但现有方法通常仍然局限于对特定吸收元素的单独建模。目前,缺乏一种能够广泛应用于多种元素的k边XANES光谱预测模型,使吸收元素方面分散的数据能够得到整合和充分利用。在这项工作中,我们通过将多元素混合数据集与局部环境的晶体拓扑方法相结合,开发了一种基于多头图注意卷积神经网络的创新“家庭桶”模型。该模型能够同时预测大范围的轻过渡金属(周期3和4)的k边XANES光谱。实验结果表明,该训练方案不仅提高了模型的精度,而且提高了模型的训练效率。在可解释性方面,获得了几个引人入胜的见解,揭示了光谱预测模型的潜在机制。我们通过采用对不同样本的一系列反应作为描述性特征来研究神经元的集体行为。值得注意的是,分析显示神经网络中的神经元表现出类似于大脑皮层中的Brodmann区域的功能分化特征。数据分析的同源性表明,不同吸收元素的样品在同一时期的相近元素之间发生了相互学习。此外,样品的注意力分数由吸收元素及其周围的原子环境共同决定。总之,本研究促进了对XANES光谱与材料结构之间关系的理解,同时为神经网络提供了有价值的见解,增强了对神经元行为的理解。
本文章由计算机程序翻译,如有差异,请以英文原文为准。
求助全文
约1分钟内获得全文 求助全文
来源期刊
Chemical Science
Chemical Science CHEMISTRY, MULTIDISCIPLINARY-
CiteScore
14.40
自引率
4.80%
发文量
1352
审稿时长
2.1 months
期刊介绍: Chemical Science is a journal that encompasses various disciplines within the chemical sciences. Its scope includes publishing ground-breaking research with significant implications for its respective field, as well as appealing to a wider audience in related areas. To be considered for publication, articles must showcase innovative and original advances in their field of study and be presented in a manner that is understandable to scientists from diverse backgrounds. However, the journal generally does not publish highly specialized research.
×
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:604180095
Book学术官方微信