A Data-Driven Construction of the Periodic Table of the Elements

By M. J. Willatt, Félix Musil, Michele Ceriotti. 2018

50 citations recorded by the source corpus.

Machine-learning of atomic-scale properties amounts to extracting correlations between structure, composition and the quantity that one wants to predict. Representing the input structure in a way that best reflects such correlations makes it possible to improve the accuracy of the model for a given amount of reference data. When using a description of the structures that is transparent and well-principled, optimizing the representation might reveal insights into the chemistry of the data set. Here we show how one can generalize the SOAP kernel to introduce a distance-dependent weight that accounts for the multi-scale nature of the interactions, and a description of correlations between chemical species. We show that this improves substantially the performance of ML models of molecular and materials stability, while making it easier to work with complex, multi-component systems and to extend SOAP to coarse-grained intermolecular potentials. The element correlations that give the best performing model show striking similarities with the conventional periodic table of the elements, providing an inspiring example of how machine learning can rediscover, and generalize, intuitive concepts that constitute the foundations of chemistry.

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Citations in the corpus

Graph Networks as a Universal Machine Learning Framework for Molecules and Crystals (2018, background): “56 Furthermore, these trends are obtained from the diverse Materials Project dataset encompassing most known crystal prototypes and 89 elements, rather than being limited to speciﬁc crystal systems 57,58 .”
DScribe: Library of Descriptors for Machine Learning in Materials Science (2019, background): “The descriptor output for multiple lo-cal sites can simply be averaged, a custom kernel can be used to combine information from multiple sites [44, 55] or the predicted property can in some cases be directly modeled as a sum of local contributions [13].”
Exploring chemical compound space with quantum-based machine learning (2019, background): “Reorganizing the periodic table represents a first and important stepping stone in that direction [71, 73].”