SchNet – A deep learning architecture for molecules and materials

Schütt, Kristof T.; Sauceda, Huziel E.; Kindermans, Pieter-Jan; Tkatchenko, Alexandre; Müller, Klaus‐Robert

doi:10.1063/1.5019779

Cited by 1,661 publications

(2,074 citation statements)

References 53 publications

Supporting

Mentioning

1,910

Contrasting

Order By: Relevance

“…Finally, graph‐based featurization has gained substantial interest in recent years. Graphs, which are natural representations for atoms (nodes) and the bonds between them (edges), have been used for molecules for many decades and have recently been applied to ML in crystals, achieving state‐of‐the‐art performance in predicting the formation energies, bandgaps, as well as metal/insulator classification …”

Section: Featurizationmentioning

confidence: 99%

A Critical Review of Machine Learning of Energy Materials

Chen

Zuo

et al. 2020

Advanced Energy Materials

399

281

View full text Add to dashboard Cite

Machine learning (ML) is rapidly revolutionizing many fields and is starting to change landscapes for physics and chemistry. With its ability to solve complex tasks autonomously, ML is being exploited as a radically new way to help find material correlations, understand materials chemistry, and accelerate the discovery of materials. Here, an in‐depth review of the application of ML to energy materials, including rechargeable alkali‐ion batteries, photovoltaics, catalysts, thermoelectrics, piezoelectrics, and superconductors, is presented. A conceptual framework is first provided for ML in materials science, with a broad overview of different ML techniques as well as best practices. This is followed by a critical discussion of how ML is applied in energy materials. This review is concluded with the perspectives on major challenges and opportunities in this exciting field.

show abstract

Section: Featurizationmentioning

confidence: 99%

A Critical Review of Machine Learning of Energy Materials

Chen

Zuo

et al. 2020

Advanced Energy Materials

399

281

View full text Add to dashboard Cite

show abstract

“…[12][13][14][15][16][17][18][19][20][21][22][23][24][25][26][27] They have been 3 used to discover materials [28][29][30][31][32][33][34][35][36][37] and study dynamical processes such as charge and exciton transfer. [38][39][40][41] Most related to this work are ML models of existing charge models, [9,[42][43][44] which are orders of magnitude faster than ab initio calculation.…”

Section: Molecular Size Training Datasetmentioning

confidence: 99%

Discovering a Transferable Charge Assignment Model Using Machine Learning

Sifain¹,

Lubbers²,

Nebgen³

et al. 2018

Preprint

View full text Add to dashboard Cite

Partial atomic charge assignment is of immense practical value to force field parametrization, molecular docking, and cheminformatics. Machine learning has emerged as a powerful tool for modeling chemistry at unprecedented computational speeds given ground-truth values, but for the task of charge assignment, the choice of ground-truth may not be obvious. In this letter, we use machine learning to discover a charge model by training a neural network to molecular dipole moments using a large, diverse set of CHNO molecular conformations. The new model, called Affordable Charge Assignment (ACA), is computationally inexpensive and predicts dipoles of out-of-sample molecules accurately. Furthermore, dipole-inferred ACA charges are transferable to dipole and even quadrupole moments of much larger molecules than those used for training. We apply ACA to long dynamical trajectories of biomolecules and successfully produce their infrared spectra. Additionally, we compare ACA with existing charge models and find that ACA assigns similar charges to Charge Model 5, but with a greatly reduced computational cost. Graphical TOC Entry Extensibility Tests Molecular Size Training DatasetKeywords machine learning, neural networks, quantum chemisty 2 Electrostatic interactions contribute strongly to the forces within and between molecules.These interactions depend on the charge density field ⇢(r), which is computationally demanding to compute. Simplified models of the charge density, such as atom-centered monopoles, are commonly employed. These partial atomic charges result in faster computation as well as provide a qualitative understanding of the underlying chemistry.[1-4] However, the decomposition of charge density into atomic charges is, by itself, an ambiguous task. Additional principles are necessary to make the charge assignment task well-defined. Here we show that a Machine Learning model, trained only on the dipole moments of small molecules, discovers a charge model that is transferable to quadrupole predictions and extensible to much larger molecules.Existing popular charge models have also been designed to reproduce observables of the electrostatic potential. The Merz-Singh-Kollman (MSK) [5,6] charge model exactly replicates the dipole moment and approximates the electrostatic potential on many points surrounding the molecule, resulting in high-quality electrostatic properties exterior to the molecule. However, MSK suffers from basis set sensitivity, particularly for "buried atoms" located inside large molecules. [7][8][9] Charge model 5 (CM5) [8] is an extension of Hirshfeld analysis, [10] with additional parametrization in order to approximately reproduce ab initio and experimental dipoles of 614 gas-phase dipoles. Unlike MSK, Hirshfeld and CM5 are nearly independent of basis set. [9] This insensitivity allows CM5 to use a single set of model parameters. The corresponding tradeoff is that its charges do not reproduce electrostatic fields as well as MSK.A limitation of these conventional charge models is that th...

show abstract

“…The neural network SchNet [10] is a variant of the earlier proposed deep tensor neural networks [8] is based on the principle of learning atom-wise representations directly from first-principles. Given the atoms of type Z 1 , .…”

Section: Schnetmentioning

confidence: 99%

“…Recently, machine learning has been successfully applied to the fast and accurate prediction of molecular properties across chemical compound space [4][5][6][7][8][9][10] and molecular dynamics simulations [11][12][13][14][15] as well as for studying properties of quantum-mechanical densities [16,17]. An indispensable ingredient to most machine learning models are molecular descriptors, which are constructed to provide an invariant, unique and efficient representation as input to machine learning models [18][19][20][21][22][23][24].…”

Section: Introductionmentioning

confidence: 99%

Capturing intensive and extensive DFT/TDDFT molecular properties with machine learning

et al. 2018

View full text Add to dashboard Cite

Abstract. Machine learning has been successfully applied to the prediction of chemical properties of small organic molecules such as energies or polarizabilities. Compared to these properties, the electronic excitation energies pose a much more challenging learning problem. Here, we examine the applicability of two existing machine learning methodologies to the prediction of excitation energies from time-dependent density functional theory. To this end, we systematically study the performance of various 2-and 3-body descriptors as well as the deep neural network SchNet to predict extensive as well as intensive properties such as the transition energies from the ground state to the first and second excited state. As perhaps expected current state-of-the-art machine learning techniques are more suited to predict extensive as opposed to intensive quantities. We speculate on the need to develop global descriptors that can describe both extensive and intensive properties on equal footing.

show abstract

SchNet – A deep learning architecture for molecules and materials

Cited by 1,661 publications

References 53 publications

A Critical Review of Machine Learning of Energy Materials

A Critical Review of Machine Learning of Energy Materials

Discovering a Transferable Charge Assignment Model Using Machine Learning

Capturing intensive and extensive DFT/TDDFT molecular properties with machine learning

Contact Info

Product

Resources

About