Subscribe to the gold package and get unlimited access to Shamra Academy
Register a new userHow to represent crystal structures for machine learning: towards fast prediction of electronic properties
488
0
0.0
(
0
)
Ask ChatGPT about the research
No Arabic abstract
High-throughput density-functional calculations of solids are extremely time consuming. As an alternative, we here propose a machine learning approach for the fast prediction of solid-state properties. To achieve this, LSDA calculations are used as training set. We focus on predicting metallic vs. insulating behavior, and on predicting the value of the density of electronic states at the Fermi energy. We find that conventional representations of the input data, such as the Coulomb matrix, are not suitable for the training of learning machines in the case of periodic solids. We propose a novel crystal structure representation for which learning and competitive prediction accuracies become possible within an unrestricted class of spd systems. Due to magnetic phenomena learning on d systems is found more difficult than in pure sp systems.
rate research
Read More
We propose an approach for exploiting machine learning to approximate electronic fields in crystalline solids subjected to deformation. Strain engineering is emerging as a widely used method for tuning the properties of materials, and this requires repeated density functional theory calculations of the unit cell subjected to strain. Repeated unit cell calculations are also required for multi-resolution studies of defects in crystalline solids. We propose an approach that uses data from such calculations to train a carefully architected machine learning approximation. We demonstrate the approach on magnesium, a promising light-weight structural material: we show that we can predict the energy and electronic fields to the level of chemical accuracy, and even capture lattice instabilities.
Lattice constants such as unit cell edge lengths and plane angles are important parameters of the periodic structures of crystal materials. Predicting crystal lattice constants has wide applications in crystal structure prediction and materials property prediction. Previous work has used machine learning models such as neural networks and support vector machines combined with composition features for lattice constant prediction and has achieved a maximum performance for cubic structures with an average $R^2$ of 0.82. Other models tailored for special materials family of a fixed form such as ABX3 perovskites can achieve much higher performance due to the homogeneity of the structures. However, these models trained with small datasets are usually not applicable to generic lattice parameter prediction of materials with diverse compositions. Herein, we report MLatticeABC, a random forest machine learning model with a new descriptor set for lattice unit cell edge length ($a,b,c$) prediction which achieves an R2 score of 0.979 for lattice parameter $a$ of cubic crystals and significant performance improvement for other crystal systems as well. Source code and trained models can be freely accessed at https://github.com/usccolumbia/MLatticeABC
Geometric information such as the space groups and crystal systems plays an important role in the properties of crystal materials. Prediction of crystal system and space group thus has wide applications in crystal material property estimation and structure prediction. Previous works on experimental X-ray diffraction (XRD) and density functional theory (DFT) based structure determination methods achieved outstanding performance, but they are not applicable for large-scale screening of materials compositions. There are also machine learning models using Magpie descriptors for composition based material space group determination, but their prediction accuracy only ranges between 0.638 and 0.907 in different kinds of crystals. Herein, we report an improved machine learning model for predicting the crystal system and space group of materials using only the formula information. Benchmark study on a dataset downloaded from Materials Project Database shows that our random forest models based on our new descriptor set, achieve significant performance improvements compared with previous work with accuracy scores ranging between 0.712 and 0.961 in terms of space group classification. Our model also shows large performance improvement for crystal system prediction. Trained models and source code are freely available at url{https://github.com/Yuxinya/SG_predict}
In this study, we present a novel approach along with the needed computational strategies for efficient and scalable feature engineering of the crystal structure in compounds of different chemical compositions. This approach utilizes a versatile and extensible framework for the quantification of a three-dimensional (3-D) voxelized crystal structure in the form of 2-point spatial correlations of multiple atomic attributes and performs principal component analysis to extract the low-dimensional features that could be used to build surrogate models for material properties of interest. An application of the proposed feature engineering framework is demonstrated on a case study involving the prediction of the formation energies of crystalline compounds using two vastly different surrogate model building strategies - local Gaussian process regression and neural networks. Specifically, it is shown that the top 25 features (i.e., principal component scores) identified by the proposed framework serve as good regressors for the formation energy of the crystalline substance for both model building strategies.
Machine learning models are increasingly used in many engineering fields thanks to the widespread digital data, growing computing power, and advanced algorithms. Artificial neural networks (ANN) is the most popular machine learning model in recent years. Although many ANN models have been used in the design and analysis of composite materials and structures, there are still some unsolved issues that hinder the acceptance of ANN models in the practical design and analysis of composite materials and structures. Moreover, the emerging machine learning techniques are posting new opportunities and challenges in the data-based design paradigm. This paper aims to give a state-of-the-art literature review of ANN models in the nonlinear constitutive modeling, multiscale surrogate modeling, and design optimization of composite materials and structures. This review has been designed to focus on the discussion of the general frameworks and benefits of ANN models to the above problems. Moreover, challenges and opportunities in each key problem are identified and discussed. This paper is expected to open the discussion of future research scope and new directions to enable efficient, robust, and accurate data-driven design and analysis of composite materials and structures.
Log in to be able to interact and post comments
comments
Fetching comments
Sign in to be able to follow your search criteria
