Article (Scientific journals)
SchNet – A deep learning architecture for molecules and materials
Schütt, Kristof T.; Sauceda, Huziel E.; Kindermans, P. J. et al.
2018In Journal of Chemical Physics, 148, p. 241722
Peer Reviewed verified by ORBi
 

Files


Full Text
122-SchNet-JCP-2018.pdf
Publisher postprint (4.2 MB)
Download

All documents in ORBilu are protected by a user license.

Send to



Details



Abstract :
[en] Deep learning has led to a paradigm shift in artificial intelligence, including web, text, and image search, speech recognition, as well as bioinformatics, with growing impact in chemical physics. Machine learning, in general, and deep learning, in particular, are ideally suitable for representing quantum-mechanical interactions, enabling us to model nonlinear potential-energy surfaces or enhancing the exploration of chemical compound space. Here we present the deep learning architecture SchNet that is specifically designed to model atomistic systems by making use of continuous-filter convolutional layers. We demonstrate the capabilities of SchNet by accurately predicting a range of properties across chemical space for molecules and materials, where our model learns chemically plausible embeddings of atom types across the periodic table. Finally, we employ SchNet to predict potential-energy surfaces and energy-conserving force fields for molecular dynamics simulations of small molecules and perform an exemplary study on the quantum-mechanical properties of C20- fullerene that would have been infeasible with regular ab initio molecular dynamics.
Disciplines :
Physics
Author, co-author :
Schütt, Kristof T.
Sauceda, Huziel E.
Kindermans, P. J.
Tkatchenko, Alexandre ;  University of Luxembourg > Faculty of Science, Technology and Communication (FSTC) > Physics and Materials Science Research Unit
Müller, Klaus R.
External co-authors :
yes
Language :
English
Title :
SchNet – A deep learning architecture for molecules and materials
Publication date :
01 March 2018
Journal title :
Journal of Chemical Physics
ISSN :
1089-7690
Publisher :
American Institute of Physics, New York, United States - New York
Volume :
148
Pages :
241722
Peer reviewed :
Peer Reviewed verified by ORBi
Focus Area :
Physics and Materials Science
Computational Sciences
Available on ORBilu :
since 02 April 2018

Statistics


Number of views
238 (9 by Unilu)
Number of downloads
3750 (12 by Unilu)

Scopus citations®
 
1026
Scopus citations®
without self-citations
977
OpenCitations
 
751
WoS citations
 
972

Bibliography


Similar publications



Contact ORBilu