Rapid and accurate predictions of perfect and defective material properties in atomistic simulation using the power of 3D CNN-based trained artificial neural networks
Abstract This article introduces an innovative approach that utilizes machine learning (ML) to address the computational challenges of accurate atomistic simulations in materials science. Focusing on the field of molecular dynamics (MD), which offers insight into material behavior at the atomic leve...
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Nature Portfolio
2024-01-01
|
| Series: | Scientific Reports |
| Online Access: | https://doi.org/10.1038/s41598-023-50893-9 |
| _version_ | 1797363342763884544 |
|---|---|
| author | Iman Peivaste Saba Ramezani Ghasem Alahyarizadeh Reza Ghaderi Ahmed Makradi Salim Belouettar |
| author_facet | Iman Peivaste Saba Ramezani Ghasem Alahyarizadeh Reza Ghaderi Ahmed Makradi Salim Belouettar |
| author_sort | Iman Peivaste |
| collection | DOAJ |
| description | Abstract This article introduces an innovative approach that utilizes machine learning (ML) to address the computational challenges of accurate atomistic simulations in materials science. Focusing on the field of molecular dynamics (MD), which offers insight into material behavior at the atomic level, the study demonstrates the potential of trained artificial neural networks (tANNs) as surrogate models. These tANNs capture complex patterns from built datasets, enabling fast and accurate predictions of material properties. The article highlights the application of 3D convolutional neural networks (CNNs) to incorporate atomistic details and defects in predictions, a significant advancement compared to current 2D image-based, or descriptor-based methods. Through a dataset of atomistic structures and MD simulations, the trained 3D CNN achieves impressive accuracy, predicting material properties with a root-mean-square error below 0.65 GPa for the prediction of elastic constants and a speed-up of approximately 185 to 2100 times compared to traditional MD simulations. This breakthrough promises to expedite materials design processes and facilitate scale-bridging in materials science, offering a new perspective on addressing computational demands in atomistic simulations. |
| first_indexed | 2024-03-08T16:20:01Z |
| format | Article |
| id | doaj.art-03fd05adaf8a4663a430b56bc846c309 |
| institution | Directory Open Access Journal |
| issn | 2045-2322 |
| language | English |
| last_indexed | 2024-03-08T16:20:01Z |
| publishDate | 2024-01-01 |
| publisher | Nature Portfolio |
| record_format | Article |
| series | Scientific Reports |
| spelling | doaj.art-03fd05adaf8a4663a430b56bc846c3092024-01-07T12:24:12ZengNature PortfolioScientific Reports2045-23222024-01-0114111410.1038/s41598-023-50893-9Rapid and accurate predictions of perfect and defective material properties in atomistic simulation using the power of 3D CNN-based trained artificial neural networksIman Peivaste0Saba Ramezani1Ghasem Alahyarizadeh2Reza Ghaderi3Ahmed Makradi4Salim Belouettar5Faculty of Engineering, Shahid Beheshti UniversityFaculty of Engineering, Shahid Beheshti UniversityFaculty of Engineering, Shahid Beheshti UniversityDepartment of Electrical Engineering, Shahid Beheshti UniversityLuxembourg Institute of Science and TechnologyLuxembourg Institute of Science and TechnologyAbstract This article introduces an innovative approach that utilizes machine learning (ML) to address the computational challenges of accurate atomistic simulations in materials science. Focusing on the field of molecular dynamics (MD), which offers insight into material behavior at the atomic level, the study demonstrates the potential of trained artificial neural networks (tANNs) as surrogate models. These tANNs capture complex patterns from built datasets, enabling fast and accurate predictions of material properties. The article highlights the application of 3D convolutional neural networks (CNNs) to incorporate atomistic details and defects in predictions, a significant advancement compared to current 2D image-based, or descriptor-based methods. Through a dataset of atomistic structures and MD simulations, the trained 3D CNN achieves impressive accuracy, predicting material properties with a root-mean-square error below 0.65 GPa for the prediction of elastic constants and a speed-up of approximately 185 to 2100 times compared to traditional MD simulations. This breakthrough promises to expedite materials design processes and facilitate scale-bridging in materials science, offering a new perspective on addressing computational demands in atomistic simulations.https://doi.org/10.1038/s41598-023-50893-9 |
| spellingShingle | Iman Peivaste Saba Ramezani Ghasem Alahyarizadeh Reza Ghaderi Ahmed Makradi Salim Belouettar Rapid and accurate predictions of perfect and defective material properties in atomistic simulation using the power of 3D CNN-based trained artificial neural networks Scientific Reports |
| title | Rapid and accurate predictions of perfect and defective material properties in atomistic simulation using the power of 3D CNN-based trained artificial neural networks |
| title_full | Rapid and accurate predictions of perfect and defective material properties in atomistic simulation using the power of 3D CNN-based trained artificial neural networks |
| title_fullStr | Rapid and accurate predictions of perfect and defective material properties in atomistic simulation using the power of 3D CNN-based trained artificial neural networks |
| title_full_unstemmed | Rapid and accurate predictions of perfect and defective material properties in atomistic simulation using the power of 3D CNN-based trained artificial neural networks |
| title_short | Rapid and accurate predictions of perfect and defective material properties in atomistic simulation using the power of 3D CNN-based trained artificial neural networks |
| title_sort | rapid and accurate predictions of perfect and defective material properties in atomistic simulation using the power of 3d cnn based trained artificial neural networks |
| url | https://doi.org/10.1038/s41598-023-50893-9 |
| work_keys_str_mv | AT imanpeivaste rapidandaccuratepredictionsofperfectanddefectivematerialpropertiesinatomisticsimulationusingthepowerof3dcnnbasedtrainedartificialneuralnetworks AT sabaramezani rapidandaccuratepredictionsofperfectanddefectivematerialpropertiesinatomisticsimulationusingthepowerof3dcnnbasedtrainedartificialneuralnetworks AT ghasemalahyarizadeh rapidandaccuratepredictionsofperfectanddefectivematerialpropertiesinatomisticsimulationusingthepowerof3dcnnbasedtrainedartificialneuralnetworks AT rezaghaderi rapidandaccuratepredictionsofperfectanddefectivematerialpropertiesinatomisticsimulationusingthepowerof3dcnnbasedtrainedartificialneuralnetworks AT ahmedmakradi rapidandaccuratepredictionsofperfectanddefectivematerialpropertiesinatomisticsimulationusingthepowerof3dcnnbasedtrainedartificialneuralnetworks AT salimbelouettar rapidandaccuratepredictionsofperfectanddefectivematerialpropertiesinatomisticsimulationusingthepowerof3dcnnbasedtrainedartificialneuralnetworks |