Deep learning generative model for crystal structure prediction
Abstract Recent advances in deep learning generative models (GMs) have created high capabilities in accessing and assessing complex high-dimensional data, allowing superior efficiency in navigating vast material configuration space in search of viable structures. Coupling such capabilities with phys...
Saved in:
| Main Authors: | , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Nature Portfolio
2024-11-01
|
| Series: | npj Computational Materials |
| Online Access: | https://doi.org/10.1038/s41524-024-01443-y |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1846165073876221952 |
|---|---|
| author | Xiaoshan Luo Zhenyu Wang Pengyue Gao Jian Lv Yanchao Wang Changfeng Chen Yanming Ma |
| author_facet | Xiaoshan Luo Zhenyu Wang Pengyue Gao Jian Lv Yanchao Wang Changfeng Chen Yanming Ma |
| author_sort | Xiaoshan Luo |
| collection | DOAJ |
| description | Abstract Recent advances in deep learning generative models (GMs) have created high capabilities in accessing and assessing complex high-dimensional data, allowing superior efficiency in navigating vast material configuration space in search of viable structures. Coupling such capabilities with physically significant data to construct trained models for materials discovery is crucial to moving this emerging field forward. Here, we present a universal GM for crystal structure prediction (CSP) via a conditional crystal diffusion variational autoencoder (Cond-CDVAE) approach, which is tailored to allow user-defined material and physical parameters such as composition and pressure. This model is trained on an expansive dataset containing over 670,000 local minimum structures, including a rich spectrum of high-pressure structures, along with ambient-pressure structures in Materials Project database. We demonstrate that the Cond-CDVAE model can generate physically plausible structures with high fidelity under diverse pressure conditions without necessitating local optimization, accurately predicting 59.3% of the 3547 unseen ambient-pressure experimental structures within 800 structure samplings, with the accuracy rate climbing to 83.2% for structures comprising fewer than 20 atoms per unit cell. These results meet or exceed those achieved via conventional CSP methods based on global optimization. The present findings showcase substantial potential of GMs in the realm of CSP. |
| format | Article |
| id | doaj-art-d12d95ee713b467d92bc271855d79d4c |
| institution | Kabale University |
| issn | 2057-3960 |
| language | English |
| publishDate | 2024-11-01 |
| publisher | Nature Portfolio |
| record_format | Article |
| series | npj Computational Materials |
| spelling | doaj-art-d12d95ee713b467d92bc271855d79d4c2024-11-17T12:38:02ZengNature Portfolionpj Computational Materials2057-39602024-11-0110111010.1038/s41524-024-01443-yDeep learning generative model for crystal structure predictionXiaoshan Luo0Zhenyu Wang1Pengyue Gao2Jian Lv3Yanchao Wang4Changfeng Chen5Yanming Ma6Key Laboratory of Material Simulation Methods and Software of Ministry of Education, College of Physics, Jilin UniversityKey Laboratory of Material Simulation Methods and Software of Ministry of Education, College of Physics, Jilin UniversityKey Laboratory of Material Simulation Methods and Software of Ministry of Education, College of Physics, Jilin UniversityKey Laboratory of Material Simulation Methods and Software of Ministry of Education, College of Physics, Jilin UniversityKey Laboratory of Material Simulation Methods and Software of Ministry of Education, College of Physics, Jilin UniversityDepartment of Physics and Astronomy, University of NevadaKey Laboratory of Material Simulation Methods and Software of Ministry of Education, College of Physics, Jilin UniversityAbstract Recent advances in deep learning generative models (GMs) have created high capabilities in accessing and assessing complex high-dimensional data, allowing superior efficiency in navigating vast material configuration space in search of viable structures. Coupling such capabilities with physically significant data to construct trained models for materials discovery is crucial to moving this emerging field forward. Here, we present a universal GM for crystal structure prediction (CSP) via a conditional crystal diffusion variational autoencoder (Cond-CDVAE) approach, which is tailored to allow user-defined material and physical parameters such as composition and pressure. This model is trained on an expansive dataset containing over 670,000 local minimum structures, including a rich spectrum of high-pressure structures, along with ambient-pressure structures in Materials Project database. We demonstrate that the Cond-CDVAE model can generate physically plausible structures with high fidelity under diverse pressure conditions without necessitating local optimization, accurately predicting 59.3% of the 3547 unseen ambient-pressure experimental structures within 800 structure samplings, with the accuracy rate climbing to 83.2% for structures comprising fewer than 20 atoms per unit cell. These results meet or exceed those achieved via conventional CSP methods based on global optimization. The present findings showcase substantial potential of GMs in the realm of CSP.https://doi.org/10.1038/s41524-024-01443-y |
| spellingShingle | Xiaoshan Luo Zhenyu Wang Pengyue Gao Jian Lv Yanchao Wang Changfeng Chen Yanming Ma Deep learning generative model for crystal structure prediction npj Computational Materials |
| title | Deep learning generative model for crystal structure prediction |
| title_full | Deep learning generative model for crystal structure prediction |
| title_fullStr | Deep learning generative model for crystal structure prediction |
| title_full_unstemmed | Deep learning generative model for crystal structure prediction |
| title_short | Deep learning generative model for crystal structure prediction |
| title_sort | deep learning generative model for crystal structure prediction |
| url | https://doi.org/10.1038/s41524-024-01443-y |
| work_keys_str_mv | AT xiaoshanluo deeplearninggenerativemodelforcrystalstructureprediction AT zhenyuwang deeplearninggenerativemodelforcrystalstructureprediction AT pengyuegao deeplearninggenerativemodelforcrystalstructureprediction AT jianlv deeplearninggenerativemodelforcrystalstructureprediction AT yanchaowang deeplearninggenerativemodelforcrystalstructureprediction AT changfengchen deeplearninggenerativemodelforcrystalstructureprediction AT yanmingma deeplearninggenerativemodelforcrystalstructureprediction |