Current browse context:
stat.ML
Change to browse by:
References & Citations
Statistics > Machine Learning
Title: Localized Coulomb Descriptors for the Gaussian Approximation Potential
(Submitted on 16 Nov 2016 (v1), last revised 6 Dec 2016 (this version, v2))
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 LC-GAP, 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 LC-GAP 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 best-performing representation has only linear dimensionality in the number of atoms in a local atomic environment, this represents an improvement both in prediction accuracy and computational cost when considered against similar Coulomb matrix-based methods.
Submission history
From: Jan Hamaekers [view email][v1] Wed, 16 Nov 2016 02:57:40 GMT (3437kb,D)
[v2] Tue, 6 Dec 2016 12:01:13 GMT (3437kb,D)
Link back to: arXiv, form interface, contact.