Gas permeability, diffusivity, and solubility in polymers: Simulation-experiment data fusion and multi-task machine learning
Abstract Machine learning (ML) models for predicting gas permeability through polymers have traditionally relied on experimental data. While these models exhibit robustness within familiar chemical domains, reliability wanes when applied to new spaces. To address this challenge, we present a multi-t...
Saved in:
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Nature Portfolio
2024-08-01
|
| Series: | npj Computational Materials |
| Online Access: | https://doi.org/10.1038/s41524-024-01373-9 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1832585608780840960 |
|---|---|
| author | Brandon K. Phan Kuan-Hsuan Shen Rishi Gurnani Huan Tran Ryan Lively Rampi Ramprasad |
| author_facet | Brandon K. Phan Kuan-Hsuan Shen Rishi Gurnani Huan Tran Ryan Lively Rampi Ramprasad |
| author_sort | Brandon K. Phan |
| collection | DOAJ |
| description | Abstract Machine learning (ML) models for predicting gas permeability through polymers have traditionally relied on experimental data. While these models exhibit robustness within familiar chemical domains, reliability wanes when applied to new spaces. To address this challenge, we present a multi-tiered multi-task learning framework empowered with advanced machine-crafted polymer fingerprinting algorithms and data fusion techniques. This framework combines scarce “high-fidelity” experimental data with abundant diverse “low-fidelity” simulation or synthetic data, resulting in predictive models that display a high level of generalizability across novel chemical spaces. Additionally, this multi-task scheme capitalizes on known physics and interrelated properties, such as gas diffusivity and solubility, both of which are closely tied to permeability. By amalgamating high throughput generated simulation data with available experimental data for gas permeability, diffusivity, and solubility for various gases, we construct multi-task deep learning models. These models can simultaneously predict all three properties for all gases under consideration, with markedly enhanced predictive accuracy, particularly compared to traditional models reliant solely on experimental data for a singular property. This strategy underscores the potential of coupling high-throughput classical simulations with data fusion methodologies to yield state-of-the-art property predictors, especially when experimental data for targeted properties is scarce. |
| format | Article |
| id | doaj-art-f919ea6941184a6190c170e11cc41716 |
| institution | Kabale University |
| issn | 2057-3960 |
| language | English |
| publishDate | 2024-08-01 |
| publisher | Nature Portfolio |
| record_format | Article |
| series | npj Computational Materials |
| spelling | doaj-art-f919ea6941184a6190c170e11cc417162025-01-26T12:43:03ZengNature Portfolionpj Computational Materials2057-39602024-08-0110111110.1038/s41524-024-01373-9Gas permeability, diffusivity, and solubility in polymers: Simulation-experiment data fusion and multi-task machine learningBrandon K. Phan0Kuan-Hsuan Shen1Rishi Gurnani2Huan Tran3Ryan Lively4Rampi Ramprasad5School of Materials Science and Engineering, Georgia Institute of TechnologySchool of Materials Science and Engineering, Georgia Institute of TechnologySchool of Materials Science and Engineering, Georgia Institute of TechnologySchool of Materials Science and Engineering, Georgia Institute of TechnologySchool of Chemical and Biomolecular Engineering, Georgia Institute of TechnologySchool of Materials Science and Engineering, Georgia Institute of TechnologyAbstract Machine learning (ML) models for predicting gas permeability through polymers have traditionally relied on experimental data. While these models exhibit robustness within familiar chemical domains, reliability wanes when applied to new spaces. To address this challenge, we present a multi-tiered multi-task learning framework empowered with advanced machine-crafted polymer fingerprinting algorithms and data fusion techniques. This framework combines scarce “high-fidelity” experimental data with abundant diverse “low-fidelity” simulation or synthetic data, resulting in predictive models that display a high level of generalizability across novel chemical spaces. Additionally, this multi-task scheme capitalizes on known physics and interrelated properties, such as gas diffusivity and solubility, both of which are closely tied to permeability. By amalgamating high throughput generated simulation data with available experimental data for gas permeability, diffusivity, and solubility for various gases, we construct multi-task deep learning models. These models can simultaneously predict all three properties for all gases under consideration, with markedly enhanced predictive accuracy, particularly compared to traditional models reliant solely on experimental data for a singular property. This strategy underscores the potential of coupling high-throughput classical simulations with data fusion methodologies to yield state-of-the-art property predictors, especially when experimental data for targeted properties is scarce.https://doi.org/10.1038/s41524-024-01373-9 |
| spellingShingle | Brandon K. Phan Kuan-Hsuan Shen Rishi Gurnani Huan Tran Ryan Lively Rampi Ramprasad Gas permeability, diffusivity, and solubility in polymers: Simulation-experiment data fusion and multi-task machine learning npj Computational Materials |
| title | Gas permeability, diffusivity, and solubility in polymers: Simulation-experiment data fusion and multi-task machine learning |
| title_full | Gas permeability, diffusivity, and solubility in polymers: Simulation-experiment data fusion and multi-task machine learning |
| title_fullStr | Gas permeability, diffusivity, and solubility in polymers: Simulation-experiment data fusion and multi-task machine learning |
| title_full_unstemmed | Gas permeability, diffusivity, and solubility in polymers: Simulation-experiment data fusion and multi-task machine learning |
| title_short | Gas permeability, diffusivity, and solubility in polymers: Simulation-experiment data fusion and multi-task machine learning |
| title_sort | gas permeability diffusivity and solubility in polymers simulation experiment data fusion and multi task machine learning |
| url | https://doi.org/10.1038/s41524-024-01373-9 |
| work_keys_str_mv | AT brandonkphan gaspermeabilitydiffusivityandsolubilityinpolymerssimulationexperimentdatafusionandmultitaskmachinelearning AT kuanhsuanshen gaspermeabilitydiffusivityandsolubilityinpolymerssimulationexperimentdatafusionandmultitaskmachinelearning AT rishigurnani gaspermeabilitydiffusivityandsolubilityinpolymerssimulationexperimentdatafusionandmultitaskmachinelearning AT huantran gaspermeabilitydiffusivityandsolubilityinpolymerssimulationexperimentdatafusionandmultitaskmachinelearning AT ryanlively gaspermeabilitydiffusivityandsolubilityinpolymerssimulationexperimentdatafusionandmultitaskmachinelearning AT rampiramprasad gaspermeabilitydiffusivityandsolubilityinpolymerssimulationexperimentdatafusionandmultitaskmachinelearning |