Rapid Identification of X-ray Diffraction Patterns Based on Very Limited Data by Interpretable Convolutional Neural Networks
Rapid Identification of X-ray Diffraction Patterns Based on Very Limited Data by Interpretable Convolutional Neural Networks
复制标题
DOI：
10.1021/acs.jcim.0c00020
复制
发表时间：
2020-03
影响因子：
5.6
通讯作者：
Hong Wang;Yunchao Xie;Dawei Li;Heng Deng;Yun-Zhi Zhao;Ming Xin;Jian Lin
中科院分区：
文献类型：
--
作者：
Hong Wang;Yunchao Xie;Dawei Li;Heng Deng;Yun-Zhi Zhao;Ming Xin;Jian Lin
来自材料表征的大量数据需要快速和自动的数据分析来加速材料发现。在本文中，我们报告了一种基于理论数据和非常有限的实验数据进行训练的卷积神经网络（CNN），用于快速识别金属有机框架（MOFs）的实验X射线衍射（XRD）图案。为了增加用于训练模型的数据，从实验中提取噪声并进行混洗，然后与从理论光谱中提取的主峰合并以合成新的光谱。首次实现了一对一的物料识别。将1012个理论MOFs模式扩展到72864个样本的整个数据集。然后将其随机洗牌并以4：1的比例分成训练（58292个样本）和验证（14572个样本）数据集。对于区分任务，优化模型在30个保留样本的测试数据集上显示出最高的前5名排名识别准确率为96.7%。对实验XRD样品的邻域成分分析（NCA）表明，来自相同材料的样品在NCA图中聚类成组。对最后一个CNN层的类激活图的分析进一步揭示了CNN模型成功地从XRD图案中识别单个MOF的机制。这种通过数据增强技术训练的CNN模型不仅将为识别不同材料的XRD图案开辟许多潜在应用，而且还为通过FTIR、拉曼和NMR等其他表征工具自主分析数据铺平了道路。
Large volumes of data from material characterizations call for rapid and automatic data analysis to accelerate materials discovery. Herein, we report a convolutional neural network (CNN) that was trained based on theoretic data and very limited experimental data for fast identification of experimental X-ray diffraction (XRD) patterns of metal-organic frameworks (MOFs). To augment the data for training the model, noise was extracted from experimental and shuffled, then merged with the main peaks that were extracted from theoretical spectra to synthesize new spectra. For the first time, one-to-one material identification was achieved. 1012 theoretic MOFs patterns were augmented to a whole dataset of 72864 samples. It was then randomly shuffled and split into training (58292 samples) and validation (14572 samples) datasets at a ratio of 4:1. For the task of discriminating, the optimized model showed the highest identification accuracy of 96.7% for the Top 5 ranking on a test dataset of 30 hold-out samples. Neighborhood components analysis (NCA) on the experimental XRD samples shows that the samples from the same material are clustered in groups in the NCA map. Analysis on the class activation maps of the last CNN layer further discloses the mechanism by which the CNN model successfully identifies individual MOFs from the XRD patterns. This CNN model trained by the data-augmentation technique would not only open numerous potential applications for identifying XRD patterns for different materials, but also pave avenues to autonomously analyze data by other characterization tools such as FTIR, Raman, and NMR.