Article (Scientific journals)
Assessment and Validation of Machine Learning Methods for Predicting Molecular Atomization Energies
Hansen, Katja; Montavon, Gregoire; Biegler, Franziska et al.
2013In JOURNAL OF CHEMICAL THEORY AND COMPUTATION, 9 (8), p. 3404-3419
Peer reviewed
 

Files


Full Text
Assessment and Validation of Machine Learning Methods for Predicting Molecular Atomization Energies.pdf
Publisher postprint (2.38 MB)
Request a copy

All documents in ORBilu are protected by a user license.

Send to



Details



Abstract :
[en] The accurate and reliable prediction of properties of molecules typically requires computationally intensive quantum-chemical calculations. Recently, machine learning techniques applied to ab initio calculations have been proposed as an efficient approach for describing the energies of molecules in their given ground-state structure throughout chemical compound space (Rupp et al. Phys. Rev. Lett. 2012 108, 058301). In this paper we outline a number of established machine learning techniques and investigate the influence of the molecular representation on the methods performance. The best methods achieve prediction errors of 3 kcal/mol for the atomization energies of a wide variety of molecules. Rationales for this performance improvement are given together with pitfalls and challenges when applying machine learning approaches to the prediction of quantum-mechanical observables.
Disciplines :
Physics
Author, co-author :
Hansen, Katja;  Fritz-Haber-Institut der Max-Planck-Gesellschaft, Berlin, Germany
Montavon, Gregoire;  Machine Learning Group, TU Berlin, Germany
Biegler, Franziska;  Machine Learning Group, TU Berlin, Germany
Siamac, Fazli;  Machine Learning Group, TU Berlin, Germany
Rupp, Matthias;  Institute of Pharmaceutical Sciences, ETH Zurich, Switzerland
Scheffler, Matthias;  Fritz-Haber-Institut der Max-Planck-Gesellschaft, Berlin, Germany
von Lilienfeld, O. Anatole;  Argonne Leadership Computing Facility, Argonne National Laboratory, Lemont, Illinois, United States
Tkatchenko, Alexandre ;  Fritz-Haber-Institut der Max-Planck-Gesellschaft, Berlin, Germany
Mueller, Klaus-Robert;  Machine Learning Group, TU Berlin, Germany ; Department of Brain and Cognitive Engineering, Korea University, Korea
External co-authors :
yes
Title :
Assessment and Validation of Machine Learning Methods for Predicting Molecular Atomization Energies
Publication date :
2013
Journal title :
JOURNAL OF CHEMICAL THEORY AND COMPUTATION
ISSN :
1549-9618
Publisher :
AMER CHEMICAL SOC, 1155 16TH ST, NW, WASHINGTON, DC 20036 USA, Unknown/unspecified
Volume :
9
Issue :
8
Pages :
3404-3419
Peer reviewed :
Peer reviewed
Commentary :
Article
Available on ORBilu :
since 17 March 2016

Statistics


Number of views
96 (2 by Unilu)
Number of downloads
0 (0 by Unilu)

Scopus citations®
 
468
Scopus citations®
without self-citations
399
OpenCitations
 
453
WoS citations
 
451

Bibliography


Similar publications



Contact ORBilu