Self-supervised Representations and Node Embedding Graph Neural Networks for Accurate and Multi-scale Analysis of Materials

Abstract

Supervised machine learning algorithms, such as graph neural networks (GNN), have successfully predicted material properties. However, the superior performance of GNN usually relies on end-to-end learning on large material datasets, which may lose the physical insight of multi-scale information about materials. And the process of labeling data consumes many resources and inevitably introduces errors, which constrains the accuracy of prediction. We propose to train the GNN model by self-supervised learning on the node and edge information of the crystal graph. Compared with the popular manually constructed material descriptors, the self-supervised atomic representation can reach better prediction performance on material properties. Furthermore, it may provide physical insights by tuning the range information. Applying the self-supervised atomic representation on the magnetic moment datasets, we show how they can extract rules and information from the magnetic materials. To incorporate rich physical information into the GNN model, we develop the node embedding graph neural networks (NEGNN) framework and show significant improvements in the prediction performance. The self-supervised material representation and the NEGNN framework may investigate in-depth information from materials and can be applied to small datasets with increased prediction accuracy.

19 pages, 12 figures, 8 tables

References in corpus (25)

• Crystal Graph Convolutional Neural Networks for an Accurate and Interpretable Prediction of Material Properties
• Fast and Accurate Modeling of Molecular Atomization Energies with Machine Learning
• SchNet - a deep learning architecture for molecules and materials
• Interpretable machine learning: definitions, methods, and applications
• E(3)-Equivariant Graph Neural Networks for Data-Efficient and Accurate Interatomic Potentials
• Graph Networks as a Universal Machine Learning Framework for Molecules and Crystals
• Machine Learning Force Fields
• Recent Advances and Applications of Deep Learning Methods in Materials Science
• Big Data of Materials Science - Critical Role of the Descriptor
• Graph Contrastive Learning with Augmentations
• Comparing molecules and solids across structural and alchemical space
• Atomistic Line Graph Neural Network for Improved Materials Property Predictions
• Nearsightedness of Electronic Matter
• Physics-inspired structural representations for molecules and materials
• High-throughput calculations of magnetic topological materials
• Interpretable and Explainable Machine Learning for Materials Science and Chemistry
• Benchmarking Materials Property Prediction Methods: The Matbench Test Set and Automatminer Reference Algorithm
• Orbital Graph Convolutional Neural Network for Material Property Prediction
• Representations of molecules and materials for interpolation of quantum-mechanical simulations via machine learning
• AtomSets -- A Hierarchical Transfer Learning Framework for Small and Large Materials Datasets
• High-throughput search for magnetic topological materials using spin-orbit spillage, machine-learning and experiments
• Structure motif centric learning framework for inorganic crystalline systems
• Benchmark for Ab Initio Prediction of Magnetic Structures based on Cluster-Multipole Theory
• Magnetic and structural studies of G-phase compound Mn${}_6$Ni${}_{16}$Si${}_7$
• Machine learning at the atomic-scale