Hybrid localized graph kernel for machine learning energy‐related properties of molecules and solids

Abstract

Nowadays, the coupling of electronic structure and machine learning techniques serves as a powerful tool to predict chemical and physical properties of a broad range of systems. With the aim of improving the accuracy of predictions, a large number of representations for molecules and solids for machine learning applications has been developed. In this work we propose a novel descriptor based on the notion of molecular graph. While graphs are largely employed in classification problems in cheminformatics or bioinformatics, they are not often used in regression problem, especially of energy‐related properties. Our method is based on a local decomposition of atomic environments and on the hybridization of two kernel functions: a graph kernel contribution that describes the chemical pattern and a Coulomb label contribution that encodes finer details of the local geometry. The accuracy of this new kernel method in energy predictions of molecular and condensed phase systems is demonstrated by considering the popular QM7 and BA10 datasets. These examples show that the hybrid localized graph kernel outperforms traditional approaches such as, for example, the smooth overlap of atomic positions and the Coulomb matrices.

Assessing the Accuracy of Machine Learning Thermodynamic Perturbation Theory: Density Functional Theory and Beyond

“…To represent the configurations within our ML model we use the smooth overlap of atomic positions (SOAP) descriptor, as implemented in the DScribe library . While several other descriptors for periodic materials have been proposed in the literature, − the choice of the SOAP descriptor provides already a satisfactory level of accuracy for MLPT applications. − …”

Assessing the Accuracy of Machine Learning Thermodynamic Perturbation Theory: Density Functional Theory and Beyond

“…To represent the configurations within our ML model we use the Smooth Overlap of Atomic Positions (SOAP) 54 descriptor, as implemented in the DScribe library 55 . While several other descriptors for periodic materials have been proposed in the literature [56][57][58] , the choice of the SOAP descriptor provides already a satisfactory level of accuracy for MLPT applications [27][28][29] .…”

Kernel methods

“…algorithms [54][55][56]. Finally, deep kernel learning [10] is a recently proposed framework that aims at surpassing the scalability problems of non-parametric kernel methods by using deep neural networks to learn highly flexible kernel functions with the possibility of training on very large datasets.…”