Optimizing transition states via kernel-based machine learning

Zachary D. Pozun; Katja Hansen; Daniel Sheppard; Matthias Rupp; Klaus Robert Müller; Graeme Henkelman

doi:10.1063/1.4707167

Optimizing transition states via kernel-based machine learning

Zachary D. Pozun, Katja Hansen, Daniel Sheppard, Matthias Rupp, Klaus Robert Müller, Graeme Henkelman

Research output: Contribution to journal › Article › peer-review

96 Citations (Scopus)

Abstract

We present a method for optimizing transition state theory dividing surfaces with support vector machines. The resulting dividing surfaces require no a priori information or intuition about reaction mechanisms. To generate optimal dividing surfaces, we apply a cycle of machine-learning and refinement of the surface by molecular dynamics sampling. We demonstrate that the machine-learned surfaces contain the relevant low-energy saddle points. The mechanisms of reactions may be extracted from the machine-learned surfaces in order to identify unexpected chemically relevant processes. Furthermore, we show that the machine-learned surfaces significantly increase the transmission coefficient for an adatom exchange involving many coupled degrees of freedom on a (100) surface when compared to a distance-based dividing surface.

Original language	English
Article number	174101
Journal	Journal of Chemical Physics
Volume	136
Issue number	17
DOIs	https://doi.org/10.1063/1.4707167
Publication status	Published - 2012 May 7

Bibliographical note

Funding Information:
All authors acknowledge support through the Institute of Pure and Applied Mathematics at UCLA. Z.D.P., D.S., and G.H. acknowledge support from the National Science Foundation (Grant No. CHE-0645497). Z.D.P. also acknowledges support from the Dorothy B. Banks Fellowship. K.H., M.R., and K.R.M. acknowledge support by the FP7-ICT programme of the European Community (PASCAL2) and Deutsche Forschungsgemeinschaft (DFG) (Grant No. MU 987/4-2). Computing resources were provided by the Texas Advanced Computing Center. This work was also supported by the World Class University Program through the National Research Foundation of Korea funded by the Ministry of Education, Science, and Technology, under Grant R31-10008.

ASJC Scopus subject areas

General Physics and Astronomy
Physical and Theoretical Chemistry

Access to Document

10.1063/1.4707167

Cite this

@article{809fcdb9a2b84721964940bef534c4cb,

title = "Optimizing transition states via kernel-based machine learning",

abstract = "We present a method for optimizing transition state theory dividing surfaces with support vector machines. The resulting dividing surfaces require no a priori information or intuition about reaction mechanisms. To generate optimal dividing surfaces, we apply a cycle of machine-learning and refinement of the surface by molecular dynamics sampling. We demonstrate that the machine-learned surfaces contain the relevant low-energy saddle points. The mechanisms of reactions may be extracted from the machine-learned surfaces in order to identify unexpected chemically relevant processes. Furthermore, we show that the machine-learned surfaces significantly increase the transmission coefficient for an adatom exchange involving many coupled degrees of freedom on a (100) surface when compared to a distance-based dividing surface.",

author = "Pozun, {Zachary D.} and Katja Hansen and Daniel Sheppard and Matthias Rupp and M{\"u}ller, {Klaus Robert} and Graeme Henkelman",

note = "Funding Information: All authors acknowledge support through the Institute of Pure and Applied Mathematics at UCLA. Z.D.P., D.S., and G.H. acknowledge support from the National Science Foundation (Grant No. CHE-0645497). Z.D.P. also acknowledges support from the Dorothy B. Banks Fellowship. K.H., M.R., and K.R.M. acknowledge support by the FP7-ICT programme of the European Community (PASCAL2) and Deutsche Forschungsgemeinschaft (DFG) (Grant No. MU 987/4-2). Computing resources were provided by the Texas Advanced Computing Center. This work was also supported by the World Class University Program through the National Research Foundation of Korea funded by the Ministry of Education, Science, and Technology, under Grant R31-10008.",

year = "2012",

month = may,

day = "7",

doi = "10.1063/1.4707167",

language = "English",

volume = "136",

journal = "Journal of Chemical Physics",

issn = "0021-9606",

publisher = "American Institute of Physics",

number = "17",

}

TY - JOUR

T1 - Optimizing transition states via kernel-based machine learning

AU - Pozun, Zachary D.

AU - Hansen, Katja

AU - Sheppard, Daniel

AU - Rupp, Matthias

AU - Müller, Klaus Robert

AU - Henkelman, Graeme

N1 - Funding Information: All authors acknowledge support through the Institute of Pure and Applied Mathematics at UCLA. Z.D.P., D.S., and G.H. acknowledge support from the National Science Foundation (Grant No. CHE-0645497). Z.D.P. also acknowledges support from the Dorothy B. Banks Fellowship. K.H., M.R., and K.R.M. acknowledge support by the FP7-ICT programme of the European Community (PASCAL2) and Deutsche Forschungsgemeinschaft (DFG) (Grant No. MU 987/4-2). Computing resources were provided by the Texas Advanced Computing Center. This work was also supported by the World Class University Program through the National Research Foundation of Korea funded by the Ministry of Education, Science, and Technology, under Grant R31-10008.

PY - 2012/5/7

Y1 - 2012/5/7

N2 - We present a method for optimizing transition state theory dividing surfaces with support vector machines. The resulting dividing surfaces require no a priori information or intuition about reaction mechanisms. To generate optimal dividing surfaces, we apply a cycle of machine-learning and refinement of the surface by molecular dynamics sampling. We demonstrate that the machine-learned surfaces contain the relevant low-energy saddle points. The mechanisms of reactions may be extracted from the machine-learned surfaces in order to identify unexpected chemically relevant processes. Furthermore, we show that the machine-learned surfaces significantly increase the transmission coefficient for an adatom exchange involving many coupled degrees of freedom on a (100) surface when compared to a distance-based dividing surface.

AB - We present a method for optimizing transition state theory dividing surfaces with support vector machines. The resulting dividing surfaces require no a priori information or intuition about reaction mechanisms. To generate optimal dividing surfaces, we apply a cycle of machine-learning and refinement of the surface by molecular dynamics sampling. We demonstrate that the machine-learned surfaces contain the relevant low-energy saddle points. The mechanisms of reactions may be extracted from the machine-learned surfaces in order to identify unexpected chemically relevant processes. Furthermore, we show that the machine-learned surfaces significantly increase the transmission coefficient for an adatom exchange involving many coupled degrees of freedom on a (100) surface when compared to a distance-based dividing surface.

UR - http://www.scopus.com/inward/record.url?scp=84862891798&partnerID=8YFLogxK

U2 - 10.1063/1.4707167

DO - 10.1063/1.4707167

M3 - Article

C2 - 22583204

AN - SCOPUS:84862891798

SN - 0021-9606

VL - 136

JO - Journal of Chemical Physics

JF - Journal of Chemical Physics

IS - 17

M1 - 174101

ER -

Optimizing transition states via kernel-based machine learning

Abstract

Bibliographical note

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this