2000 character limit reached
CTGNN: Crystal Transformer Graph Neural Network for Crystal Material Property Prediction (2405.11502v1)
Published 19 May 2024 in cond-mat.mtrl-sci and physics.comp-ph
Abstract: The combination of deep learning algorithm and materials science has made significant progress in predicting novel materials and understanding various behaviours of materials. Here, we introduced a new model called as the Crystal Transformer Graph Neural Network (CTGNN), which combines the advantages of Transformer model and graph neural networks to address the complexity of structure-properties relation of material data. Compared to the state-of-the-art models, CTGNN incorporates the graph network structure for capturing local atomic interactions and the dual-Transformer structures to model intra-crystal and inter-atomic relationships comprehensively. The benchmark carried on by the proposed CTGNN indicates that CTGNN significantly outperforms existing models like CGCNN and MEGNET in the prediction of formation energy and bandgap properties. Our work highlights the potential of CTGNN to enhance the performance of properties prediction and accelerates the discovery of new materials, particularly for perovskite materials.
- Graph neural networks: A review of methods and applications. AI open, 1:57–81, 2020.
- Highly accurate protein structure prediction with alphafold. Nature, 596(7873):583–589, 2021.
- Pushing the frontiers of density functionals by solving the fractional electron problem. Science, 374(6573):1385–1389, 2021.
- Advancing mathematics by guiding human intuition with ai. Nature, 600(7887):70–74, 2021.
- Crystal graph convolutional neural networks for an accurate and interpretable prediction of material properties. Physical review letters, 120(14):145301, 2018.
- Crystal diffusion variational autoencoder for periodic material generation. arXiv preprint arXiv:2110.06197, 2021.
- Predicting the band gaps of inorganic solids by machine learning. The Journal of Physical Chemistry Letters, 9(7):1668–1673, 2018. PMID: 29532658.
- Machine learning for molecular and materials science. Nature, 559(7715):547–555, 2018. Published 2018/07/01.
- Deep learning. nature, 521(7553):436–444, 2015.
- A comprehensive survey on graph neural networks. IEEE transactions on neural networks and learning systems, 32(1):4–24, 2020.
- Graph networks as a universal machine learning framework for molecules and crystals. Chemistry of Materials, 31(9):3564–3572, 2019.
- Atomistic line graph neural network for improved materials property predictions. npj Computational Materials, 7(1):185, 2021.
- Developing an improved crystal graph convolutional neural network framework for accelerated materials discovery. Physical Review Materials, 4(6):063801, 2020.
- Orbnet: Deep learning for quantum chemistry using symmetry-adapted atomic-orbital features. The Journal of chemical physics, 153(12), 2020.
- Directional message passing for molecular graphs. arXiv preprint arXiv:2003.03123, 2020.
- Fast and uncertainty-aware directional message passing for non-equilibrium molecules. arXiv preprint arXiv:2011.14115, 2020.
- Heterogeneous molecular graph neural networks for predicting molecule properties. In 2020 IEEE International Conference on Data Mining (ICDM), pages 492–500. IEEE, 2020.
- Quantum-chemical insights from deep tensor neural networks. Nature communications, 8(1):13890, 2017.
- Cormorant: Covariant molecular neural networks. Advances in neural information processing systems, 32, 2019.
- Molecular mechanics-driven graph neural network with multiplex graph for molecular structures. arXiv preprint arXiv:2011.07457, 2020.
- Schnetpack: A deep learning toolbox for atomistic systems. Journal of chemical theory and computation, 15(1):448–455, 2018.
- Elemnet: Deep learning the chemistry of materials from only elemental composition. Scientific reports, 8(1):17593, 2018.
- Combining schnet and sharc: The schnarc machine learning approach for excited-state dynamics. The journal of physical chemistry letters, 11(10):3828–3834, 2020.
- Bondnet: a graph neural network for the prediction of bond dissociation energies for charged molecules. Chemical science, 12(5):1858–1868, 2021.
- Universal fragment descriptors for predicting properties of inorganic crystals. Nature communications, 8(1):15679, 2017.
- Learning properties of ordered and disordered materials from multi-fidelity data. Nature Computational Science, 1(1):46–53, 2021.
- Accelerated discovery of stable lead-free hybrid organic-inorganic perovskites via machine learning. Nature communications, 9(1):3405, 2018.
- Automating crystal-structure phase mapping by combining deep learning with constraint reasoning. Nature Machine Intelligence, 3(9):812–822, 2021.
- Fast inverse design of microstructures via generative invariance networks. Nature Computational Science, 1(3):229–238, 2021.
- Attention is all you need. Advances in neural information processing systems, 30, 2017.
- Periodic graph transformers for crystal material property prediction. Advances in Neural Information Processing Systems, 35:15066–15080, 2022.
- Transformer for graphs: An overview from architecture perspective. arXiv preprint arXiv:2202.08455, 2022.
- Ada-gnn: Atom-distance-angle graph neural network for crystal material property prediction. https://synthical.com/article/2e48f156-8b9d-4f18-a52a-3e730cdd1ede, 0 2024.
- Tg-gnn: Transformer based geometric enhancement graph neural network for molecular property prediction, 2022.
- Global attention based graph convolutional neural networks for improved materials property prediction. ArXiv, abs/2003.13379, 2020.
- Kamal Choudhary. Jarvis-dft, 2014.
- A hybrid organic-inorganic perovskite dataset. Scientific Data, 4(1):1–11, 2017.
- Discovery of pb-free perovskite solar cells via high-throughput simulation on the k computer. The journal of physical chemistry letters, 8(19):4826–4831, 2017.
Paper Prompts
Sign up for free to create and run prompts on this paper using GPT-5.
Top Community Prompts
Collections
Sign up for free to add this paper to one or more collections.
