What problem does this paper attempt to address?

The paper aims to address the significant challenge of predicting crystal structures with ideal properties in materials science research. Specifically, the research objectives include: 1. **Large-scale generation of hypothetical structures**: Using the graph theory-assisted universal structure searcher (MAGUS) to generate over 3.4 million hypothetical structures from 17,277 compositions (involving 63 elements from the periodic table) in the Open Quantum Materials Database (OQMD). 2. **Rapid optimization of a large number of hypothetical structures**: Employing a pre-trained universal interatomic potential model—the Crystal Hamiltonian Graph Neural Network (CHGNet)—to rapidly optimize these numerous hypothetical structures. 3. **Screening thermodynamically stable structures**: Using Density Functional Theory (DFT) validation to identify 4,145 successfully optimized structures, and further identifying 2,368 structures with energies lower than those in the original OQMD. 4. **Dynamic stability assessment**: Using CHGNet to identify 647 dynamically stable structures, and validating through DFT that 123 out of 200 randomly selected structures are indeed dynamically stable, with a success rate of 61.5%. 5. **Prediction of physical properties**: Using Graph Neural Network (GNN) models to predict the lattice thermal conductivity (LTC) and heat capacity of the remaining structures, thereby screening for structures with ultra-low LTC and high heat capacity, suitable for advanced energy conversion and storage applications. In summary, this research proposes a method combining graph theory and a pre-trained universal interatomic potential model to accelerate the discovery of new thermodynamically and dynamically stable structures with target material properties, such as low thermal conductivity and high heat capacity, which are crucial for energy conversion and storage applications. This approach overcomes the limitations of traditional methods in terms of computational cost and time consumption, demonstrating potential in large-scale new structure searches.

Graph theory and graph neural network assisted high-throughput crystal structure prediction and screening for energy conversion and storage

Crystal structure prediction by combining graph network and optimization algorithm

MAGUS: machine learning and graph theory assisted universal structure searcher

Graph deep learning accelerated efficient crystal structure search and feature extraction

CrysToGraph: A Comprehensive Predictive Model for Crystal Materials Properties and the Benchmark

Enhancing crystal structure prediction by combining computational and experimental data via graph networks

Enhancing Crystal Structure Prediction by decomposition methods based on graph theory

AlphaCrystal: Contact map based crystal structure prediction using deep learning

Graph Transformer Networks for Accurate Band Structure Prediction: An End-to-End Approach

Correction: Graph theory and graph neural network assisted high-throughput crystal structure prediction and screening for energy conversion and storage

Study of crystal properties based on attention mechanism and crystal graph convolutional neural network

CTGNN: Crystal Transformer Graph Neural Network for Crystal Material Property Prediction

Defect graph neural networks for materials discovery in high-temperature clean-energy applications

Enhancing Crystal Structure Prediction by Decomposition and Evolution Schemes Based on Graph Theory

Large-scale machine-learning-assisted exploration of the whole materials space

Tripartite interaction representation algorithm for crystal graph neural networks

A geometric-information-enhanced crystal graph network for predicting properties of materials

Structure prediction and materials design with generative neural networks

Novel machine learning framework for thermal conductivity prediction by crystal graph convolution embedded ensemble

Crystal Graph Convolutional Neural Networks for an Accurate and Interpretable Prediction of Material Properties

Improving Crystal Property Prediction from a Multiplex Graph Perspective