Learning Representations of Molecules and Materials with Atomistic Neural Networks

Kristof T. Schütt, Alexandre Tkatchenko, Klaus Robert Müller

Research output: Chapter in Book/Report/Conference proceedingChapter

Abstract

Deep learning has been shown to learn efficient representations for structured data such as images, text, or audio. In this chapter, we present neural network architectures that are able to learn efficient representations of molecules and materials. In particular, the continuous-filter convolutional network SchNet accurately predicts chemical properties across compositional and configurational space on a variety of datasets. Beyond that, we analyze the obtained representations to find evidence that their spatial and chemical properties agree with chemical intuition.

Original languageEnglish
Title of host publicationLecture Notes in Physics
PublisherSpringer
Pages215-230
Number of pages16
DOIs
Publication statusPublished - 2020

Publication series

NameLecture Notes in Physics
Volume968
ISSN (Print)0075-8450
ISSN (Electronic)1616-6361

ASJC Scopus subject areas

  • Physics and Astronomy (miscellaneous)

Fingerprint Dive into the research topics of 'Learning Representations of Molecules and Materials with Atomistic Neural Networks'. Together they form a unique fingerprint.

  • Cite this

    Schütt, K. T., Tkatchenko, A., & Müller, K. R. (2020). Learning Representations of Molecules and Materials with Atomistic Neural Networks. In Lecture Notes in Physics (pp. 215-230). (Lecture Notes in Physics; Vol. 968). Springer. https://doi.org/10.1007/978-3-030-40245-7_11