Algorithms for Finding Saddle Points and Minimum Energy Paths Using Gaussian Process Regression
This doctoral dissertation has been conducted under a convention for the joint supervision at Aalto University (Finland) and University of Iceland (Iceland). Chemical reactions and other transitions involving rearrangements of atoms can be studied theoretically by analyzing a potential energy surfac...
Published in: | Nanosystems: Physics, Chemistry, Mathematics |
---|---|
Main Author: | |
Other Authors: | , , , , , , , , |
Format: | Doctoral or Postdoctoral Thesis |
Language: | English |
Published: |
Aalto University
2019
|
Subjects: | |
Online Access: | https://aaltodoc.aalto.fi/handle/123456789/41794 |
id |
ftaaltouniv:oai:aaltodoc.aalto.fi:123456789/41794 |
---|---|
record_format |
openpolar |
institution |
Open Polar |
collection |
Aalto University Publication Archive (Aaltodoc) |
op_collection_id |
ftaaltouniv |
language |
English |
topic |
Computer science saddle point minimum energy path Gaussian process machine learning satulapiste minimienergiapolku gaussinen prosessi koneoppiminen |
spellingShingle |
Computer science saddle point minimum energy path Gaussian process machine learning satulapiste minimienergiapolku gaussinen prosessi koneoppiminen Koistinen, Olli-Pekka Algorithms for Finding Saddle Points and Minimum Energy Paths Using Gaussian Process Regression |
topic_facet |
Computer science saddle point minimum energy path Gaussian process machine learning satulapiste minimienergiapolku gaussinen prosessi koneoppiminen |
description |
This doctoral dissertation has been conducted under a convention for the joint supervision at Aalto University (Finland) and University of Iceland (Iceland). Chemical reactions and other transitions involving rearrangements of atoms can be studied theoretically by analyzing a potential energy surface defined in a high-dimensional space of atom coordinates. Local minimum points of the energy surface correspond to stable states of the system, and minimum energy paths connecting these states characterize mechanisms of possible transitions. Of particular interest is often the maximum point of the minimum energy path, which is located at a first-order saddle point of the energy surface and can be used to estimate the activation energy and rate of the particular transition. Minimum energy paths and saddle points between two known states have been traditionally searched with iterative methods where a chain of discrete points of the coordinate space is moved and stretched towards a minimum energy path according to imaginary forces based on gradient vectors of the potential energy surface. The actual saddle point can be found by reversing the component of the gradient vector parallel to the path at one of the points of the chain and letting this point climb along the path towards the saddle point. If the end state of the transition is unknown, the saddle point can be searched correspondingly by rotating a pair of closely spaced points towards the orientation of the lowest curvature, reversing the gradient component corresponding to this direction, and moving the pair towards the saddle point. These methods may, however, require hundreds of iterations, and since accurate evaluation of the gradient vector is often computationally expensive, the information obtained from previous iterations should be utilized as efficiently as possible to decrease the number of iterations. Using statistical models, an approximation to the energy surface can be constructed, and a minimum energy path or a saddle point can be searched on the ... |
author2 |
Perustieteiden korkeakoulu School of Science Tietotekniikan laitos Department of Computer Science Vehtari, Aki, Prof., Aalto University, Department of Computer Science, Finland Jónsson, Hannes, Prof., University of Iceland, Iceland University of Iceland Aalto-yliopisto Aalto University |
format |
Doctoral or Postdoctoral Thesis |
author |
Koistinen, Olli-Pekka |
author_facet |
Koistinen, Olli-Pekka |
author_sort |
Koistinen, Olli-Pekka |
title |
Algorithms for Finding Saddle Points and Minimum Energy Paths Using Gaussian Process Regression |
title_short |
Algorithms for Finding Saddle Points and Minimum Energy Paths Using Gaussian Process Regression |
title_full |
Algorithms for Finding Saddle Points and Minimum Energy Paths Using Gaussian Process Regression |
title_fullStr |
Algorithms for Finding Saddle Points and Minimum Energy Paths Using Gaussian Process Regression |
title_full_unstemmed |
Algorithms for Finding Saddle Points and Minimum Energy Paths Using Gaussian Process Regression |
title_sort |
algorithms for finding saddle points and minimum energy paths using gaussian process regression |
publisher |
Aalto University |
publishDate |
2019 |
url |
https://aaltodoc.aalto.fi/handle/123456789/41794 |
long_lat |
ENVELOPE(73.483,73.483,-53.017,-53.017) |
geographic |
Saddle Point |
geographic_facet |
Saddle Point |
genre |
Iceland |
genre_facet |
Iceland |
op_relation |
Aalto University publication series DOCTORAL DISSERTATIONS 222/2019 [Publication 1]: Olli-Pekka Koistinen, Emile Maras, Aki Vehtari, and Hannes Jónsson. Minimum energy path calculations with Gaussian process regression. Nanosystems: Physics, Chemistry, Mathematics, volume 7, issue 6, pages 925–935, December 2016. DOI:10.17586/2220-8054-2016-7-6-925-935 [Publication 2]: Olli-Pekka Koistinen, Freyja B. Dagbjartsdóttir, Vilhjálmur Ásgeirsson, Aki Vehtari, and Hannes Jónsson. Nudged elastic band calculations accelerated with Gaussian process regression. The Journal of Chemical Physics, volume 147, issue 15, article 152720, 14 pages, September 2017. DOI:10.1063/1.4986787 [Publication 3]: Olli-Pekka Koistinen, Vilhjálmur Ásgeirsson, Aki Vehtari, and Hannes Jónsson. Nudged elastic band calculations accelerated with Gaussian process regression based on inverse interatomic distances. Journal of Chemical Theory and Computation, volume 15, issue 12, pages 6738–6751, October 2019. DOI:10.1021/acs.jctc.9b00692 [Publication 4]: Olli-Pekka Koistinen, Vilhjálmur Ásgeirsson, Aki Vehtari, and Hannes Jónsson. Minimum mode saddle point searches using Gaussian process regression with inverse-distance covariance function. Accepted for publication in Journal of Chemical Theory and Computation, 20 pages, December 2019. DOI:10.1021/acs.jctc.9b01038 [Errata file]: Errata Olli-Pekka Koistinen DD-222/2019 publications P1 and P2 978-952-60-8851-8 (electronic) 978-952-60-8850-1 (printed) 1799-4942 (electronic) 1799-4934 (printed) 1799-4934 (ISSN-L) https://aaltodoc.aalto.fi/handle/123456789/41794 URN:ISBN:978-952-60-8851-8 |
op_doi |
https://doi.org/10.17586/2220-8054-2016-7-6-925-935 https://doi.org/10.1063/1.4986787 https://doi.org/10.1021/acs.jctc.9b00692 https://doi.org/10.1021/acs.jctc.9b01038 |
container_title |
Nanosystems: Physics, Chemistry, Mathematics |
container_start_page |
925 |
op_container_end_page |
935 |
_version_ |
1766039752876752896 |
spelling |
ftaaltouniv:oai:aaltodoc.aalto.fi:123456789/41794 2023-05-15T16:49:37+02:00 Algorithms for Finding Saddle Points and Minimum Energy Paths Using Gaussian Process Regression Gaussisia prosesseja hyödyntäviä menetelmiä satulapisteiden ja minimienergiapolkujen etsintään Koistinen, Olli-Pekka Perustieteiden korkeakoulu School of Science Tietotekniikan laitos Department of Computer Science Vehtari, Aki, Prof., Aalto University, Department of Computer Science, Finland Jónsson, Hannes, Prof., University of Iceland, Iceland University of Iceland Aalto-yliopisto Aalto University 2019 application/pdf https://aaltodoc.aalto.fi/handle/123456789/41794 en eng Aalto University Aalto-yliopisto Aalto University publication series DOCTORAL DISSERTATIONS 222/2019 [Publication 1]: Olli-Pekka Koistinen, Emile Maras, Aki Vehtari, and Hannes Jónsson. Minimum energy path calculations with Gaussian process regression. Nanosystems: Physics, Chemistry, Mathematics, volume 7, issue 6, pages 925–935, December 2016. DOI:10.17586/2220-8054-2016-7-6-925-935 [Publication 2]: Olli-Pekka Koistinen, Freyja B. Dagbjartsdóttir, Vilhjálmur Ásgeirsson, Aki Vehtari, and Hannes Jónsson. Nudged elastic band calculations accelerated with Gaussian process regression. The Journal of Chemical Physics, volume 147, issue 15, article 152720, 14 pages, September 2017. DOI:10.1063/1.4986787 [Publication 3]: Olli-Pekka Koistinen, Vilhjálmur Ásgeirsson, Aki Vehtari, and Hannes Jónsson. Nudged elastic band calculations accelerated with Gaussian process regression based on inverse interatomic distances. Journal of Chemical Theory and Computation, volume 15, issue 12, pages 6738–6751, October 2019. DOI:10.1021/acs.jctc.9b00692 [Publication 4]: Olli-Pekka Koistinen, Vilhjálmur Ásgeirsson, Aki Vehtari, and Hannes Jónsson. Minimum mode saddle point searches using Gaussian process regression with inverse-distance covariance function. Accepted for publication in Journal of Chemical Theory and Computation, 20 pages, December 2019. DOI:10.1021/acs.jctc.9b01038 [Errata file]: Errata Olli-Pekka Koistinen DD-222/2019 publications P1 and P2 978-952-60-8851-8 (electronic) 978-952-60-8850-1 (printed) 1799-4942 (electronic) 1799-4934 (printed) 1799-4934 (ISSN-L) https://aaltodoc.aalto.fi/handle/123456789/41794 URN:ISBN:978-952-60-8851-8 Computer science saddle point minimum energy path Gaussian process machine learning satulapiste minimienergiapolku gaussinen prosessi koneoppiminen G5 Artikkeliväitöskirja text Doctoral dissertation (article-based) Väitöskirja (artikkeli) 2019 ftaaltouniv https://doi.org/10.17586/2220-8054-2016-7-6-925-935 https://doi.org/10.1063/1.4986787 https://doi.org/10.1021/acs.jctc.9b00692 https://doi.org/10.1021/acs.jctc.9b01038 2022-12-15T19:19:35Z This doctoral dissertation has been conducted under a convention for the joint supervision at Aalto University (Finland) and University of Iceland (Iceland). Chemical reactions and other transitions involving rearrangements of atoms can be studied theoretically by analyzing a potential energy surface defined in a high-dimensional space of atom coordinates. Local minimum points of the energy surface correspond to stable states of the system, and minimum energy paths connecting these states characterize mechanisms of possible transitions. Of particular interest is often the maximum point of the minimum energy path, which is located at a first-order saddle point of the energy surface and can be used to estimate the activation energy and rate of the particular transition. Minimum energy paths and saddle points between two known states have been traditionally searched with iterative methods where a chain of discrete points of the coordinate space is moved and stretched towards a minimum energy path according to imaginary forces based on gradient vectors of the potential energy surface. The actual saddle point can be found by reversing the component of the gradient vector parallel to the path at one of the points of the chain and letting this point climb along the path towards the saddle point. If the end state of the transition is unknown, the saddle point can be searched correspondingly by rotating a pair of closely spaced points towards the orientation of the lowest curvature, reversing the gradient component corresponding to this direction, and moving the pair towards the saddle point. These methods may, however, require hundreds of iterations, and since accurate evaluation of the gradient vector is often computationally expensive, the information obtained from previous iterations should be utilized as efficiently as possible to decrease the number of iterations. Using statistical models, an approximation to the energy surface can be constructed, and a minimum energy path or a saddle point can be searched on the ... Doctoral or Postdoctoral Thesis Iceland Aalto University Publication Archive (Aaltodoc) Saddle Point ENVELOPE(73.483,73.483,-53.017,-53.017) Nanosystems: Physics, Chemistry, Mathematics 925 935 |