Equivariant message passing for the prediction of tensorial properties and molecular spectra

Kristof Schütt, Oliver Unke, Michael Gastegger
Proceedings of the 38th International Conference on Machine Learning, PMLR 139:9377-9388, 2021.

Abstract

Message passing neural networks have become a method of choice for learning on graphs, in particular the prediction of chemical properties and the acceleration of molecular dynamics studies. While they readily scale to large training data sets, previous approaches have proven to be less data efficient than kernel methods. We identify limitations of invariant representations as a major reason and extend the message passing formulation to rotationally equivariant representations. On this basis, we propose the polarizable atom interaction neural network (PaiNN) and improve on common molecule benchmarks over previous networks, while reducing model size and inference time. We leverage the equivariant atomwise representations obtained by PaiNN for the prediction of tensorial properties. Finally, we apply this to the simulation of molecular spectra, achieving speedups of 4-5 orders of magnitude compared to the electronic structure reference.

Cite this Paper

BibTeX
@InProceedings{pmlr-v139-schutt21a, title = {Equivariant message passing for the prediction of tensorial properties and molecular spectra}, author = {Sch{\"u}tt, Kristof and Unke, Oliver and Gastegger, Michael}, booktitle = {Proceedings of the 38th International Conference on Machine Learning}, pages = {9377--9388}, year = {2021}, editor = {Meila, Marina and Zhang, Tong}, volume = {139}, series = {Proceedings of Machine Learning Research}, month = {18--24 Jul}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v139/schutt21a/schutt21a.pdf}, url = {https://proceedings.mlr.press/v139/schutt21a.html}, abstract = {Message passing neural networks have become a method of choice for learning on graphs, in particular the prediction of chemical properties and the acceleration of molecular dynamics studies. While they readily scale to large training data sets, previous approaches have proven to be less data efficient than kernel methods. We identify limitations of invariant representations as a major reason and extend the message passing formulation to rotationally equivariant representations. On this basis, we propose the polarizable atom interaction neural network (PaiNN) and improve on common molecule benchmarks over previous networks, while reducing model size and inference time. We leverage the equivariant atomwise representations obtained by PaiNN for the prediction of tensorial properties. Finally, we apply this to the simulation of molecular spectra, achieving speedups of 4-5 orders of magnitude compared to the electronic structure reference.} }
Endnote
%0 Conference Paper %T Equivariant message passing for the prediction of tensorial properties and molecular spectra %A Kristof Schütt %A Oliver Unke %A Michael Gastegger %B Proceedings of the 38th International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2021 %E Marina Meila %E Tong Zhang %F pmlr-v139-schutt21a %I PMLR %P 9377--9388 %U https://proceedings.mlr.press/v139/schutt21a.html %V 139 %X Message passing neural networks have become a method of choice for learning on graphs, in particular the prediction of chemical properties and the acceleration of molecular dynamics studies. While they readily scale to large training data sets, previous approaches have proven to be less data efficient than kernel methods. We identify limitations of invariant representations as a major reason and extend the message passing formulation to rotationally equivariant representations. On this basis, we propose the polarizable atom interaction neural network (PaiNN) and improve on common molecule benchmarks over previous networks, while reducing model size and inference time. We leverage the equivariant atomwise representations obtained by PaiNN for the prediction of tensorial properties. Finally, we apply this to the simulation of molecular spectra, achieving speedups of 4-5 orders of magnitude compared to the electronic structure reference.
APA
Schütt, K., Unke, O. & Gastegger, M.. (2021). Equivariant message passing for the prediction of tensorial properties and molecular spectra. Proceedings of the 38th International Conference on Machine Learning, in Proceedings of Machine Learning Research 139:9377-9388 Available from https://proceedings.mlr.press/v139/schutt21a.html.

Related Material