Neural Message Passing with Edge Updates for Predicting Properties of Molecules and Materials

Research output: Contribution to conferencePaper – Annual report year: 2019Research

Standard

Neural Message Passing with Edge Updates for Predicting Properties of Molecules and Materials. / Jørgensen, Peter Bjørn; Jacobsen, Karsten Wedel; Schmidt, Mikkel Nørgaard.

2018. Paper presented at 32nd Conference on Neural Information Processing Systems, Montreal, Canada.

Research output: Contribution to conferencePaper – Annual report year: 2019Research

Harvard

Jørgensen, PB, Jacobsen, KW & Schmidt, MN 2018, 'Neural Message Passing with Edge Updates for Predicting Properties of Molecules and Materials' Paper presented at 32nd Conference on Neural Information Processing Systems, Montreal, Canada, 02/12/2018 - 08/12/2018, .

APA

Jørgensen, P. B., Jacobsen, K. W., & Schmidt, M. N. (2018). Neural Message Passing with Edge Updates for Predicting Properties of Molecules and Materials. Paper presented at 32nd Conference on Neural Information Processing Systems, Montreal, Canada.

CBE

Jørgensen PB, Jacobsen KW, Schmidt MN. 2018. Neural Message Passing with Edge Updates for Predicting Properties of Molecules and Materials. Paper presented at 32nd Conference on Neural Information Processing Systems, Montreal, Canada.

MLA

Vancouver

Jørgensen PB, Jacobsen KW, Schmidt MN. Neural Message Passing with Edge Updates for Predicting Properties of Molecules and Materials. 2018. Paper presented at 32nd Conference on Neural Information Processing Systems, Montreal, Canada.

Author

Bibtex

@conference{31bef22ac7034baca72d1f08d3b16c4b,
title = "Neural Message Passing with Edge Updates for Predicting Properties of Molecules and Materials",
abstract = "Neural message passing on molecular graphs is one of the most promising methods for predicting formation energy and other properties of molecules and materials. In this work we extend the neural message passing model with an edge update network which allows the information exchanged between atoms to depend on the hidden state of the receiving atom. We benchmark the proposed model on three publicly available datasets (QM9, The Materials Project and OQMD) and show that the proposed model yields superior prediction of formation energies and other properties on all three datasets in comparison with the best published results. Furthermore we investigate different methods for constructing the graph used to represent crystalline structures and we find that using a graph based on K-nearest neighbors achieves better prediction accuracy than using maximum distance cutoff or the Voronoi tessellation graph.",
author = "J{\o}rgensen, {Peter Bj{\o}rn} and Jacobsen, {Karsten Wedel} and Schmidt, {Mikkel N{\o}rgaard}",
year = "2018",
language = "English",
note = "32nd Conference on Neural Information Processing Systems, NIPS 2018 ; Conference date: 02-12-2018 Through 08-12-2018",

}

RIS

TY - CONF

T1 - Neural Message Passing with Edge Updates for Predicting Properties of Molecules and Materials

AU - Jørgensen, Peter Bjørn

AU - Jacobsen, Karsten Wedel

AU - Schmidt, Mikkel Nørgaard

PY - 2018

Y1 - 2018

N2 - Neural message passing on molecular graphs is one of the most promising methods for predicting formation energy and other properties of molecules and materials. In this work we extend the neural message passing model with an edge update network which allows the information exchanged between atoms to depend on the hidden state of the receiving atom. We benchmark the proposed model on three publicly available datasets (QM9, The Materials Project and OQMD) and show that the proposed model yields superior prediction of formation energies and other properties on all three datasets in comparison with the best published results. Furthermore we investigate different methods for constructing the graph used to represent crystalline structures and we find that using a graph based on K-nearest neighbors achieves better prediction accuracy than using maximum distance cutoff or the Voronoi tessellation graph.

AB - Neural message passing on molecular graphs is one of the most promising methods for predicting formation energy and other properties of molecules and materials. In this work we extend the neural message passing model with an edge update network which allows the information exchanged between atoms to depend on the hidden state of the receiving atom. We benchmark the proposed model on three publicly available datasets (QM9, The Materials Project and OQMD) and show that the proposed model yields superior prediction of formation energies and other properties on all three datasets in comparison with the best published results. Furthermore we investigate different methods for constructing the graph used to represent crystalline structures and we find that using a graph based on K-nearest neighbors achieves better prediction accuracy than using maximum distance cutoff or the Voronoi tessellation graph.

M3 - Paper

ER -