Deep energy-pressure regression for a thermodynamically consistent EOS model
In this paper, we aim to explore novel machine learning (ML) techniques to facilitate and accelerate the construction of universal equation-Of-State (EOS) models with a high accuracy while ensuring important thermodynamic consistency. When applying ML to fit a universal EOS model, there are two key...
| Main Authors: | , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IOP Publishing
2024-01-01
|
| Series: | Machine Learning: Science and Technology |
| Subjects: | |
| Online Access: | https://doi.org/10.1088/2632-2153/ad2626 |
| _version_ | 1797302272880803840 |
|---|---|
| author | Dayou Yu Deep Shankar Pandey Joshua Hinz Deyan Mihaylov Valentin V Karasiev S X Hu Qi Yu |
| author_facet | Dayou Yu Deep Shankar Pandey Joshua Hinz Deyan Mihaylov Valentin V Karasiev S X Hu Qi Yu |
| author_sort | Dayou Yu |
| collection | DOAJ |
| description | In this paper, we aim to explore novel machine learning (ML) techniques to facilitate and accelerate the construction of universal equation-Of-State (EOS) models with a high accuracy while ensuring important thermodynamic consistency. When applying ML to fit a universal EOS model, there are two key requirements: (1) a high prediction accuracy to ensure precise estimation of relevant physics properties and (2) physical interpretability to support important physics-related downstream applications. We first identify a set of fundamental challenges from the accuracy perspective, including an extremely wide range of input/output space and highly sparse training data. We demonstrate that while a neural network (NN) model may fit the EOS data well, the black-box nature makes it difficult to provide physically interpretable results, leading to weak accountability of prediction results outside the training range and lack of guarantee to meet important thermodynamic consistency constraints. To this end, we propose a principled deep regression model that can be trained following a meta-learning style to predict the desired quantities with a high accuracy using scarce training data. We further introduce a uniquely designed kernel-based regularizer for accurate uncertainty quantification. An ensemble technique is leveraged to battle model overfitting with improved prediction stability. Auto-differentiation is conducted to verify that necessary thermodynamic consistency conditions are maintained. Our evaluation results show an excellent fit of the EOS table and the predicted values are ready to use for important physics-related tasks. |
| first_indexed | 2024-03-07T23:34:25Z |
| format | Article |
| id | doaj.art-5271bdf845104c05b3198a5c50122220 |
| institution | Directory Open Access Journal |
| issn | 2632-2153 |
| language | English |
| last_indexed | 2024-03-07T23:34:25Z |
| publishDate | 2024-01-01 |
| publisher | IOP Publishing |
| record_format | Article |
| series | Machine Learning: Science and Technology |
| spelling | doaj.art-5271bdf845104c05b3198a5c501222202024-02-20T09:58:52ZengIOP PublishingMachine Learning: Science and Technology2632-21532024-01-015101503110.1088/2632-2153/ad2626Deep energy-pressure regression for a thermodynamically consistent EOS modelDayou Yu0https://orcid.org/0009-0002-2373-4907Deep Shankar Pandey1https://orcid.org/0009-0006-1404-3716Joshua Hinz2Deyan Mihaylov3https://orcid.org/0000-0002-8874-5503Valentin V Karasiev4https://orcid.org/0000-0003-3445-6797S X Hu5https://orcid.org/0000-0003-2465-3818Qi Yu6https://orcid.org/0000-0002-0426-5407Rochester Institute Of Technology , Rochester, NY, United States of AmericaRochester Institute Of Technology , Rochester, NY, United States of AmericaLaboratory for Laser Energetics, University of Rochester , Rochester, NY, United States of AmericaLaboratory for Laser Energetics, University of Rochester , Rochester, NY, United States of AmericaLaboratory for Laser Energetics, University of Rochester , Rochester, NY, United States of AmericaLaboratory for Laser Energetics, University of Rochester , Rochester, NY, United States of AmericaRochester Institute Of Technology , Rochester, NY, United States of AmericaIn this paper, we aim to explore novel machine learning (ML) techniques to facilitate and accelerate the construction of universal equation-Of-State (EOS) models with a high accuracy while ensuring important thermodynamic consistency. When applying ML to fit a universal EOS model, there are two key requirements: (1) a high prediction accuracy to ensure precise estimation of relevant physics properties and (2) physical interpretability to support important physics-related downstream applications. We first identify a set of fundamental challenges from the accuracy perspective, including an extremely wide range of input/output space and highly sparse training data. We demonstrate that while a neural network (NN) model may fit the EOS data well, the black-box nature makes it difficult to provide physically interpretable results, leading to weak accountability of prediction results outside the training range and lack of guarantee to meet important thermodynamic consistency constraints. To this end, we propose a principled deep regression model that can be trained following a meta-learning style to predict the desired quantities with a high accuracy using scarce training data. We further introduce a uniquely designed kernel-based regularizer for accurate uncertainty quantification. An ensemble technique is leveraged to battle model overfitting with improved prediction stability. Auto-differentiation is conducted to verify that necessary thermodynamic consistency conditions are maintained. Our evaluation results show an excellent fit of the EOS table and the predicted values are ready to use for important physics-related tasks.https://doi.org/10.1088/2632-2153/ad2626equation-of-statemeta-learninguncertainty quantification |
| spellingShingle | Dayou Yu Deep Shankar Pandey Joshua Hinz Deyan Mihaylov Valentin V Karasiev S X Hu Qi Yu Deep energy-pressure regression for a thermodynamically consistent EOS model Machine Learning: Science and Technology equation-of-state meta-learning uncertainty quantification |
| title | Deep energy-pressure regression for a thermodynamically consistent EOS model |
| title_full | Deep energy-pressure regression for a thermodynamically consistent EOS model |
| title_fullStr | Deep energy-pressure regression for a thermodynamically consistent EOS model |
| title_full_unstemmed | Deep energy-pressure regression for a thermodynamically consistent EOS model |
| title_short | Deep energy-pressure regression for a thermodynamically consistent EOS model |
| title_sort | deep energy pressure regression for a thermodynamically consistent eos model |
| topic | equation-of-state meta-learning uncertainty quantification |
| url | https://doi.org/10.1088/2632-2153/ad2626 |
| work_keys_str_mv | AT dayouyu deepenergypressureregressionforathermodynamicallyconsistenteosmodel AT deepshankarpandey deepenergypressureregressionforathermodynamicallyconsistenteosmodel AT joshuahinz deepenergypressureregressionforathermodynamicallyconsistenteosmodel AT deyanmihaylov deepenergypressureregressionforathermodynamicallyconsistenteosmodel AT valentinvkarasiev deepenergypressureregressionforathermodynamicallyconsistenteosmodel AT sxhu deepenergypressureregressionforathermodynamicallyconsistenteosmodel AT qiyu deepenergypressureregressionforathermodynamicallyconsistenteosmodel |