An invertible, invariant crystal representation for inverse design of solid-state materials using generative deep learning
Abstract The past decade has witnessed rapid progress in deep learning for molecular design, owing to the availability of invertible and invariant representations for molecules such as simplified molecular-input line-entry system (SMILES), which has powered cheminformatics since the late 1980s. Howe...
| Main Authors: | , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Nature Portfolio
2023-11-01
|
| Series: | Nature Communications |
| Online Access: | https://doi.org/10.1038/s41467-023-42870-7 |
| _version_ | 1797636793513803776 |
|---|---|
| author | Hang Xiao Rong Li Xiaoyang Shi Yan Chen Liangliang Zhu Xi Chen Lei Wang |
| author_facet | Hang Xiao Rong Li Xiaoyang Shi Yan Chen Liangliang Zhu Xi Chen Lei Wang |
| author_sort | Hang Xiao |
| collection | DOAJ |
| description | Abstract The past decade has witnessed rapid progress in deep learning for molecular design, owing to the availability of invertible and invariant representations for molecules such as simplified molecular-input line-entry system (SMILES), which has powered cheminformatics since the late 1980s. However, the design of elemental components and their structural arrangement in solid-state materials to achieve certain desired properties is still a long-standing challenge in physics, chemistry and biology. This is primarily due to, unlike molecular inverse design, the lack of an invertible crystal representation that satisfies translational, rotational, and permutational invariances. To address this issue, we have developed a simplified line-input crystal-encoding system (SLICES), which is a string-based crystal representation that satisfies both invertibility and invariances. The reconstruction routine of SLICES successfully reconstructed 94.95% of over 40,000 structurally and chemically diverse crystal structures, showcasing an unprecedented invertibility. Furthermore, by only encoding compositional and topological data, SLICES guarantees invariances. We demonstrate the application of SLICES in the inverse design of direct narrow-gap semiconductors for optoelectronic applications. As a string-based, invertible, and invariant crystal representation, SLICES shows promise as a useful tool for in silico materials discovery. |
| first_indexed | 2024-03-11T12:39:15Z |
| format | Article |
| id | doaj.art-02ae383390d34261830312b7871f6aab |
| institution | Directory Open Access Journal |
| issn | 2041-1723 |
| language | English |
| last_indexed | 2024-03-11T12:39:15Z |
| publishDate | 2023-11-01 |
| publisher | Nature Portfolio |
| record_format | Article |
| series | Nature Communications |
| spelling | doaj.art-02ae383390d34261830312b7871f6aab2023-11-05T12:22:37ZengNature PortfolioNature Communications2041-17232023-11-0114111210.1038/s41467-023-42870-7An invertible, invariant crystal representation for inverse design of solid-state materials using generative deep learningHang Xiao0Rong Li1Xiaoyang Shi2Yan Chen3Liangliang Zhu4Xi Chen5Lei Wang6School of Interdisciplinary Studies, Lingnan UniversitySchool of Chemical Engineering, Northwest UniversityDepartment of Environmental and Sustainable Engineering, State University of New York at AlbanyLaboratory for Multiscale Mechanics and Medical Science, SV LAB, School of Aerospace, Xi’an Jiaotong UniversitySchool of Chemical Engineering, Northwest UniversitySchool of Interdisciplinary Studies, Lingnan UniversityNational Laboratory of Solid-State Microstructures, School of Physics, Nanjing UniversityAbstract The past decade has witnessed rapid progress in deep learning for molecular design, owing to the availability of invertible and invariant representations for molecules such as simplified molecular-input line-entry system (SMILES), which has powered cheminformatics since the late 1980s. However, the design of elemental components and their structural arrangement in solid-state materials to achieve certain desired properties is still a long-standing challenge in physics, chemistry and biology. This is primarily due to, unlike molecular inverse design, the lack of an invertible crystal representation that satisfies translational, rotational, and permutational invariances. To address this issue, we have developed a simplified line-input crystal-encoding system (SLICES), which is a string-based crystal representation that satisfies both invertibility and invariances. The reconstruction routine of SLICES successfully reconstructed 94.95% of over 40,000 structurally and chemically diverse crystal structures, showcasing an unprecedented invertibility. Furthermore, by only encoding compositional and topological data, SLICES guarantees invariances. We demonstrate the application of SLICES in the inverse design of direct narrow-gap semiconductors for optoelectronic applications. As a string-based, invertible, and invariant crystal representation, SLICES shows promise as a useful tool for in silico materials discovery.https://doi.org/10.1038/s41467-023-42870-7 |
| spellingShingle | Hang Xiao Rong Li Xiaoyang Shi Yan Chen Liangliang Zhu Xi Chen Lei Wang An invertible, invariant crystal representation for inverse design of solid-state materials using generative deep learning Nature Communications |
| title | An invertible, invariant crystal representation for inverse design of solid-state materials using generative deep learning |
| title_full | An invertible, invariant crystal representation for inverse design of solid-state materials using generative deep learning |
| title_fullStr | An invertible, invariant crystal representation for inverse design of solid-state materials using generative deep learning |
| title_full_unstemmed | An invertible, invariant crystal representation for inverse design of solid-state materials using generative deep learning |
| title_short | An invertible, invariant crystal representation for inverse design of solid-state materials using generative deep learning |
| title_sort | invertible invariant crystal representation for inverse design of solid state materials using generative deep learning |
| url | https://doi.org/10.1038/s41467-023-42870-7 |
| work_keys_str_mv | AT hangxiao aninvertibleinvariantcrystalrepresentationforinversedesignofsolidstatematerialsusinggenerativedeeplearning AT rongli aninvertibleinvariantcrystalrepresentationforinversedesignofsolidstatematerialsusinggenerativedeeplearning AT xiaoyangshi aninvertibleinvariantcrystalrepresentationforinversedesignofsolidstatematerialsusinggenerativedeeplearning AT yanchen aninvertibleinvariantcrystalrepresentationforinversedesignofsolidstatematerialsusinggenerativedeeplearning AT liangliangzhu aninvertibleinvariantcrystalrepresentationforinversedesignofsolidstatematerialsusinggenerativedeeplearning AT xichen aninvertibleinvariantcrystalrepresentationforinversedesignofsolidstatematerialsusinggenerativedeeplearning AT leiwang aninvertibleinvariantcrystalrepresentationforinversedesignofsolidstatematerialsusinggenerativedeeplearning AT hangxiao invertibleinvariantcrystalrepresentationforinversedesignofsolidstatematerialsusinggenerativedeeplearning AT rongli invertibleinvariantcrystalrepresentationforinversedesignofsolidstatematerialsusinggenerativedeeplearning AT xiaoyangshi invertibleinvariantcrystalrepresentationforinversedesignofsolidstatematerialsusinggenerativedeeplearning AT yanchen invertibleinvariantcrystalrepresentationforinversedesignofsolidstatematerialsusinggenerativedeeplearning AT liangliangzhu invertibleinvariantcrystalrepresentationforinversedesignofsolidstatematerialsusinggenerativedeeplearning AT xichen invertibleinvariantcrystalrepresentationforinversedesignofsolidstatematerialsusinggenerativedeeplearning AT leiwang invertibleinvariantcrystalrepresentationforinversedesignofsolidstatematerialsusinggenerativedeeplearning |