Optimal criteria and their asymptotic form for data selection in data-driven reduced-order modelling with Gaussian process regression
We derive criteria for the selection of datapoints used for data-driven reduced-order modelling and other areas of supervised learning based on Gaussian process regression (GPR). While this is a well-studied area in the fields of active learning and optimal experimental design, most criteria in the...
| Main Authors: | , |
|---|---|
| Other Authors: | |
| Format: | Article |
| Language: | English |
| Published: |
The Royal Society
2024
|
| Subjects: | |
| Online Access: | https://hdl.handle.net/1721.1/154218 |
| _version_ | 1811072287746555904 |
|---|---|
| author | Sapsis, Themistoklis P. Blanchard, Antoine |
| author2 | Massachusetts Institute of Technology. Department of Mechanical Engineering |
| author_facet | Massachusetts Institute of Technology. Department of Mechanical Engineering Sapsis, Themistoklis P. Blanchard, Antoine |
| author_sort | Sapsis, Themistoklis P. |
| collection | MIT |
| description | We derive criteria for the selection of datapoints used for data-driven reduced-order modelling and other areas of supervised learning based on Gaussian process regression (GPR). While this is a well-studied area in the fields of active learning and optimal experimental design, most criteria in the literature are empirical. Here we introduce an optimality condition for the selection of a new input defined as the minimizer of the distance between the approximated output probability density function (pdf) of the reduced-order model and the exact one. Given that the exact pdf is unknown, we define the selection criterion as the supremum over the unit sphere of the native Hilbert space for the GPR. The resulting selection criterion, however, has a form that is difficult to compute. We combine results from GPR theory and asymptotic analysis to derive a computable form of the defined optimality criterion that is valid in the limit of small predictive variance. The derived asymptotic form of the selection criterion leads to convergence of the GPR model that guarantees a balanced distribution of data resources between probable and large-deviation outputs, resulting in an effective way of sampling towards data-driven reduced-order modelling. |
| first_indexed | 2024-09-23T09:03:35Z |
| format | Article |
| id | mit-1721.1/154218 |
| institution | Massachusetts Institute of Technology |
| language | English |
| last_indexed | 2024-09-23T09:03:35Z |
| publishDate | 2024 |
| publisher | The Royal Society |
| record_format | dspace |
| spelling | mit-1721.1/1542182024-09-19T15:30:44Z Optimal criteria and their asymptotic form for data selection in data-driven reduced-order modelling with Gaussian process regression Sapsis, Themistoklis P. Blanchard, Antoine Massachusetts Institute of Technology. Department of Mechanical Engineering General Physics and Astronomy General Engineering General Mathematics We derive criteria for the selection of datapoints used for data-driven reduced-order modelling and other areas of supervised learning based on Gaussian process regression (GPR). While this is a well-studied area in the fields of active learning and optimal experimental design, most criteria in the literature are empirical. Here we introduce an optimality condition for the selection of a new input defined as the minimizer of the distance between the approximated output probability density function (pdf) of the reduced-order model and the exact one. Given that the exact pdf is unknown, we define the selection criterion as the supremum over the unit sphere of the native Hilbert space for the GPR. The resulting selection criterion, however, has a form that is difficult to compute. We combine results from GPR theory and asymptotic analysis to derive a computable form of the defined optimality criterion that is valid in the limit of small predictive variance. The derived asymptotic form of the selection criterion leads to convergence of the GPR model that guarantees a balanced distribution of data resources between probable and large-deviation outputs, resulting in an effective way of sampling towards data-driven reduced-order modelling. 2024-04-18T17:38:13Z 2024-04-18T17:38:13Z 2022-06-20 2024-04-18T17:34:39Z Article http://purl.org/eprint/type/JournalArticle 1364-503X 1471-2962 https://hdl.handle.net/1721.1/154218 Sapsis, Themistoklis P. and Blanchard, Antoine. 2022. "Optimal criteria and their asymptotic form for data selection in data-driven reduced-order modelling with Gaussian process regression." Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences, 380 (2229). en 10.1098/rsta.2021.0197 Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences Creative Commons Attribution-Noncommercial-ShareAlike http://creativecommons.org/licenses/by-nc-sa/4.0/ application/pdf The Royal Society arxiv |
| spellingShingle | General Physics and Astronomy General Engineering General Mathematics Sapsis, Themistoklis P. Blanchard, Antoine Optimal criteria and their asymptotic form for data selection in data-driven reduced-order modelling with Gaussian process regression |
| title | Optimal criteria and their asymptotic form for data selection in data-driven reduced-order modelling with Gaussian process regression |
| title_full | Optimal criteria and their asymptotic form for data selection in data-driven reduced-order modelling with Gaussian process regression |
| title_fullStr | Optimal criteria and their asymptotic form for data selection in data-driven reduced-order modelling with Gaussian process regression |
| title_full_unstemmed | Optimal criteria and their asymptotic form for data selection in data-driven reduced-order modelling with Gaussian process regression |
| title_short | Optimal criteria and their asymptotic form for data selection in data-driven reduced-order modelling with Gaussian process regression |
| title_sort | optimal criteria and their asymptotic form for data selection in data driven reduced order modelling with gaussian process regression |
| topic | General Physics and Astronomy General Engineering General Mathematics |
| url | https://hdl.handle.net/1721.1/154218 |
| work_keys_str_mv | AT sapsisthemistoklisp optimalcriteriaandtheirasymptoticformfordataselectionindatadrivenreducedordermodellingwithgaussianprocessregression AT blanchardantoine optimalcriteriaandtheirasymptoticformfordataselectionindatadrivenreducedordermodellingwithgaussianprocessregression |