A generative deep learning framework for inverse design of compositionally complex bulk metallic glasses
Abstract The design of bulk metallic glasses (BMGs) via machine learning (ML) has been a topic of active research recently. However, the prior ML models were mostly built upon supervised learning algorithms with human inputs to navigate the high dimensional compositional space, which becomes ineffic...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Nature Portfolio
2023-01-01
|
| Series: | npj Computational Materials |
| Online Access: | https://doi.org/10.1038/s41524-023-00968-y |
| _version_ | 1811175765932244992 |
|---|---|
| author | Ziqing Zhou Yinghui Shang Xiaodi Liu Yong Yang |
| author_facet | Ziqing Zhou Yinghui Shang Xiaodi Liu Yong Yang |
| author_sort | Ziqing Zhou |
| collection | DOAJ |
| description | Abstract The design of bulk metallic glasses (BMGs) via machine learning (ML) has been a topic of active research recently. However, the prior ML models were mostly built upon supervised learning algorithms with human inputs to navigate the high dimensional compositional space, which becomes inefficient with the increasing compositional complexity in BMGs. Here, we develop a generative deep-learning framework to directly generate compositionally complex BMGs, such as high entropy BMGs. Our framework is built on the unsupervised Generative Adversarial Network (GAN) algorithm for data generation and the supervised Boosted Trees algorithm for data evaluation. We studied systematically the confounding effect of various data descriptors and the literature data on the effectiveness of our framework both numerically and experimentally. Most importantly, we demonstrate that our generative deep learning framework is capable of producing composition-property mappings, therefore paving the way for the inverse design of BMGs. |
| first_indexed | 2024-04-10T19:42:16Z |
| format | Article |
| id | doaj.art-acb88a2212714f2e999f041ab97ee3da |
| institution | Directory Open Access Journal |
| issn | 2057-3960 |
| language | English |
| last_indexed | 2024-04-10T19:42:16Z |
| publishDate | 2023-01-01 |
| publisher | Nature Portfolio |
| record_format | Article |
| series | npj Computational Materials |
| spelling | doaj.art-acb88a2212714f2e999f041ab97ee3da2023-01-29T12:17:29ZengNature Portfolionpj Computational Materials2057-39602023-01-01911810.1038/s41524-023-00968-yA generative deep learning framework for inverse design of compositionally complex bulk metallic glassesZiqing Zhou0Yinghui Shang1Xiaodi Liu2Yong Yang3Department of Mechanical Engineering, College of Engineering, City University of Hong Kong, Kowloon TongDepartment of Mechanical Engineering, College of Engineering, City University of Hong Kong, Kowloon TongCollege of Mechatronics and Control Engineering, Shenzhen UniversityDepartment of Mechanical Engineering, College of Engineering, City University of Hong Kong, Kowloon TongAbstract The design of bulk metallic glasses (BMGs) via machine learning (ML) has been a topic of active research recently. However, the prior ML models were mostly built upon supervised learning algorithms with human inputs to navigate the high dimensional compositional space, which becomes inefficient with the increasing compositional complexity in BMGs. Here, we develop a generative deep-learning framework to directly generate compositionally complex BMGs, such as high entropy BMGs. Our framework is built on the unsupervised Generative Adversarial Network (GAN) algorithm for data generation and the supervised Boosted Trees algorithm for data evaluation. We studied systematically the confounding effect of various data descriptors and the literature data on the effectiveness of our framework both numerically and experimentally. Most importantly, we demonstrate that our generative deep learning framework is capable of producing composition-property mappings, therefore paving the way for the inverse design of BMGs.https://doi.org/10.1038/s41524-023-00968-y |
| spellingShingle | Ziqing Zhou Yinghui Shang Xiaodi Liu Yong Yang A generative deep learning framework for inverse design of compositionally complex bulk metallic glasses npj Computational Materials |
| title | A generative deep learning framework for inverse design of compositionally complex bulk metallic glasses |
| title_full | A generative deep learning framework for inverse design of compositionally complex bulk metallic glasses |
| title_fullStr | A generative deep learning framework for inverse design of compositionally complex bulk metallic glasses |
| title_full_unstemmed | A generative deep learning framework for inverse design of compositionally complex bulk metallic glasses |
| title_short | A generative deep learning framework for inverse design of compositionally complex bulk metallic glasses |
| title_sort | generative deep learning framework for inverse design of compositionally complex bulk metallic glasses |
| url | https://doi.org/10.1038/s41524-023-00968-y |
| work_keys_str_mv | AT ziqingzhou agenerativedeeplearningframeworkforinversedesignofcompositionallycomplexbulkmetallicglasses AT yinghuishang agenerativedeeplearningframeworkforinversedesignofcompositionallycomplexbulkmetallicglasses AT xiaodiliu agenerativedeeplearningframeworkforinversedesignofcompositionallycomplexbulkmetallicglasses AT yongyang agenerativedeeplearningframeworkforinversedesignofcompositionallycomplexbulkmetallicglasses AT ziqingzhou generativedeeplearningframeworkforinversedesignofcompositionallycomplexbulkmetallicglasses AT yinghuishang generativedeeplearningframeworkforinversedesignofcompositionallycomplexbulkmetallicglasses AT xiaodiliu generativedeeplearningframeworkforinversedesignofcompositionallycomplexbulkmetallicglasses AT yongyang generativedeeplearningframeworkforinversedesignofcompositionallycomplexbulkmetallicglasses |