Active learning approach to simulations of strongly correlated matter with the ghost Gutzwiller approximation
Quantum embedding (QE) methods such as the ghost Gutzwiller approximation (gGA) offer a powerful approach to simulating strongly correlated systems, but come with the computational bottleneck of computing the ground state of an auxiliary embedding Hamiltonian (EH) iteratively. In this work, we intro...
| Main Authors: | , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
American Physical Society
2024-03-01
|
| Series: | Physical Review Research |
| Online Access: | http://doi.org/10.1103/PhysRevResearch.6.013242 |
| _version_ | 1797210216343797760 |
|---|---|
| author | Marius S. Frank Denis G. Artiukhin Tsung-Han Lee Yongxin Yao Kipton Barros Ove Christiansen Nicola Lanatà |
| author_facet | Marius S. Frank Denis G. Artiukhin Tsung-Han Lee Yongxin Yao Kipton Barros Ove Christiansen Nicola Lanatà |
| author_sort | Marius S. Frank |
| collection | DOAJ |
| description | Quantum embedding (QE) methods such as the ghost Gutzwiller approximation (gGA) offer a powerful approach to simulating strongly correlated systems, but come with the computational bottleneck of computing the ground state of an auxiliary embedding Hamiltonian (EH) iteratively. In this work, we introduce an active learning (AL) framework integrated within the gGA to address this challenge. The methodology is applied to the single-band Hubbard model and results in a significant reduction in the number of instances where the EH must be solved. Through a principal component analysis (PCA), we find that the EH parameters form a low-dimensional structure that is largely independent of the geometric specifics of the systems, especially in the strongly correlated regime. Our AL strategy enables us to discover this low-dimensionality structure on the fly, while leveraging it for reducing the computational cost of gGA, laying the groundwork for more efficient simulations of complex strongly correlated materials. |
| first_indexed | 2024-04-24T10:07:04Z |
| format | Article |
| id | doaj.art-31f83d3e2b9c4728b90e3bbe749a28f7 |
| institution | Directory Open Access Journal |
| issn | 2643-1564 |
| language | English |
| last_indexed | 2024-04-24T10:07:04Z |
| publishDate | 2024-03-01 |
| publisher | American Physical Society |
| record_format | Article |
| series | Physical Review Research |
| spelling | doaj.art-31f83d3e2b9c4728b90e3bbe749a28f72024-04-12T17:39:59ZengAmerican Physical SocietyPhysical Review Research2643-15642024-03-016101324210.1103/PhysRevResearch.6.013242Active learning approach to simulations of strongly correlated matter with the ghost Gutzwiller approximationMarius S. FrankDenis G. ArtiukhinTsung-Han LeeYongxin YaoKipton BarrosOve ChristiansenNicola LanatàQuantum embedding (QE) methods such as the ghost Gutzwiller approximation (gGA) offer a powerful approach to simulating strongly correlated systems, but come with the computational bottleneck of computing the ground state of an auxiliary embedding Hamiltonian (EH) iteratively. In this work, we introduce an active learning (AL) framework integrated within the gGA to address this challenge. The methodology is applied to the single-band Hubbard model and results in a significant reduction in the number of instances where the EH must be solved. Through a principal component analysis (PCA), we find that the EH parameters form a low-dimensional structure that is largely independent of the geometric specifics of the systems, especially in the strongly correlated regime. Our AL strategy enables us to discover this low-dimensionality structure on the fly, while leveraging it for reducing the computational cost of gGA, laying the groundwork for more efficient simulations of complex strongly correlated materials.http://doi.org/10.1103/PhysRevResearch.6.013242 |
| spellingShingle | Marius S. Frank Denis G. Artiukhin Tsung-Han Lee Yongxin Yao Kipton Barros Ove Christiansen Nicola Lanatà Active learning approach to simulations of strongly correlated matter with the ghost Gutzwiller approximation Physical Review Research |
| title | Active learning approach to simulations of strongly correlated matter with the ghost Gutzwiller approximation |
| title_full | Active learning approach to simulations of strongly correlated matter with the ghost Gutzwiller approximation |
| title_fullStr | Active learning approach to simulations of strongly correlated matter with the ghost Gutzwiller approximation |
| title_full_unstemmed | Active learning approach to simulations of strongly correlated matter with the ghost Gutzwiller approximation |
| title_short | Active learning approach to simulations of strongly correlated matter with the ghost Gutzwiller approximation |
| title_sort | active learning approach to simulations of strongly correlated matter with the ghost gutzwiller approximation |
| url | http://doi.org/10.1103/PhysRevResearch.6.013242 |
| work_keys_str_mv | AT mariussfrank activelearningapproachtosimulationsofstronglycorrelatedmatterwiththeghostgutzwillerapproximation AT denisgartiukhin activelearningapproachtosimulationsofstronglycorrelatedmatterwiththeghostgutzwillerapproximation AT tsunghanlee activelearningapproachtosimulationsofstronglycorrelatedmatterwiththeghostgutzwillerapproximation AT yongxinyao activelearningapproachtosimulationsofstronglycorrelatedmatterwiththeghostgutzwillerapproximation AT kiptonbarros activelearningapproachtosimulationsofstronglycorrelatedmatterwiththeghostgutzwillerapproximation AT ovechristiansen activelearningapproachtosimulationsofstronglycorrelatedmatterwiththeghostgutzwillerapproximation AT nicolalanata activelearningapproachtosimulationsofstronglycorrelatedmatterwiththeghostgutzwillerapproximation |