De novo exploration and self-guided learning of potential-energy surfaces
Interatomic potential models based on machine learning (ML) are rapidly developing as tools for material simulations. However, because of their flexibility, they require large fitting databases that are normally created with substantial manual selection and tuning of reference configurations. Here,...
| Main Authors: | , , |
|---|---|
| Format: | Journal article |
| Language: | English |
| Published: |
Nature Research
2019
|
| _version_ | 1797100947513540608 |
|---|---|
| author | Bernstein, N Csanyi, G Volker L Deringer |
| author_facet | Bernstein, N Csanyi, G Volker L Deringer |
| author_sort | Bernstein, N |
| collection | OXFORD |
| description | Interatomic potential models based on machine learning (ML) are rapidly developing as tools for material simulations. However, because of their flexibility, they require large fitting databases that are normally created with substantial manual selection and tuning of reference configurations. Here, we show that ML potentials can be built in a largely automated fashion, exploring and fitting potential-energy surfaces from the beginning (de novo) within one and the same protocol. The key enabling step is the use of a configuration-averaged kernel metric that allows one to select the few most relevant and diverse structures at each step. The resulting potentials are accurate and robust for the wide range of configurations that occur during structure searching, despite only requiring a relatively small number of single-point DFT calculations on small unit cells. We apply the method to materials with diverse chemical nature and coordination environments, marking an important step toward the more routine application of ML potentials in physics, chemistry, and materials science. |
| first_indexed | 2024-03-07T05:44:57Z |
| format | Journal article |
| id | oxford-uuid:e6eb4713-ba09-4328-9ba2-5cc72cca6ae0 |
| institution | University of Oxford |
| language | English |
| last_indexed | 2024-03-07T05:44:57Z |
| publishDate | 2019 |
| publisher | Nature Research |
| record_format | dspace |
| spelling | oxford-uuid:e6eb4713-ba09-4328-9ba2-5cc72cca6ae02022-03-27T10:34:30ZDe novo exploration and self-guided learning of potential-energy surfacesJournal articlehttp://purl.org/coar/resource_type/c_dcae04bcuuid:e6eb4713-ba09-4328-9ba2-5cc72cca6ae0EnglishSymplectic ElementsNature Research2019Bernstein, NCsanyi, GVolker L DeringerInteratomic potential models based on machine learning (ML) are rapidly developing as tools for material simulations. However, because of their flexibility, they require large fitting databases that are normally created with substantial manual selection and tuning of reference configurations. Here, we show that ML potentials can be built in a largely automated fashion, exploring and fitting potential-energy surfaces from the beginning (de novo) within one and the same protocol. The key enabling step is the use of a configuration-averaged kernel metric that allows one to select the few most relevant and diverse structures at each step. The resulting potentials are accurate and robust for the wide range of configurations that occur during structure searching, despite only requiring a relatively small number of single-point DFT calculations on small unit cells. We apply the method to materials with diverse chemical nature and coordination environments, marking an important step toward the more routine application of ML potentials in physics, chemistry, and materials science. |
| spellingShingle | Bernstein, N Csanyi, G Volker L Deringer De novo exploration and self-guided learning of potential-energy surfaces |
| title | De novo exploration and self-guided learning of potential-energy surfaces |
| title_full | De novo exploration and self-guided learning of potential-energy surfaces |
| title_fullStr | De novo exploration and self-guided learning of potential-energy surfaces |
| title_full_unstemmed | De novo exploration and self-guided learning of potential-energy surfaces |
| title_short | De novo exploration and self-guided learning of potential-energy surfaces |
| title_sort | de novo exploration and self guided learning of potential energy surfaces |
| work_keys_str_mv | AT bernsteinn denovoexplorationandselfguidedlearningofpotentialenergysurfaces AT csanyig denovoexplorationandselfguidedlearningofpotentialenergysurfaces AT volkerlderinger denovoexplorationandselfguidedlearningofpotentialenergysurfaces |