Publica
Hier finden Sie wissenschaftliche Publikationen aus den FraunhoferInstituten. LCGAP: Localized Coulomb Descriptors for the Gaussian Approximation Potential
 Griebel, Michael; Schüller, Anton; Schweitzer, Marc Alexander: Scientific Computing and Algorithms in Industrial Simulations : Projects and Products of Fraunhofer SCAI Cham: Springer International Publishing, 2017 ISBN: 9783319624570 (Print) ISBN: 9783319624587 (Online) pp.2542 

 English 
 Book Article 
 Fraunhofer SCAI () 
Abstract
We introduce a novel class of localized atomic environment representations based upon the Coulomb matrix. By combining these functions with the Gaussian approximation potential approach, we present LCGAP, a new system for generating atomic potentials through machine learning (ML). Tests on the QM7, QM7b and GDB9 biomolecular datasets demonstrate that potentials created with LCGAP can successfully predict atomization energies for molecules larger than those used for training to chemical accuracy, and can (in the case of QM7b) also be used to predict a range of other atomic properties with accuracy in line with the recent literature. As the bestperforming representation has only linear dimensionality in the number of atoms in a local atomic environment, this represents an improvement in both prediction accuracy and computational cost when compared to similar Coulomb matrixbased methods.