Recent Advances and Applications of Graph Convolution Neural Network Methods in Materials Science

With the development of artificial intelligence (AI), AI plus science is increasingly valued, presenting new perspectives to scientific research. The research on using machine learning (including deep learning) to discover patterns from data and predict targeted material properties has received widespread attention, which will have a profound impact in material science studies. In recent years, there has been an increased interest in the use of deep learning in materials science, which has led to significant progress in both fundamental and applied research. One of the most notable advancements is the development of graph convolutional neural network models, which combine graph neural networks and convolutional neural networks to achieve outstanding results in materials science and bridge effectively the deep learning models and material properties predictions. The availability of large materials databases due to the rise of big data has further enhanced the relevance of these models in the field. We present, in this article, a comprehensive overview of graph convolutional neural network models, explaining their fundamental principles and highlighting a few examples of their applications in materials science, as well as current trends. The limitations and challenges that these models face, as well as the potential for future research in this dynamic area are also discussed.