GemNet: Universal Directional Graph Neural Networks for Molecules

This page is about our papers

GemNet: Universal Directional Graph Neural Networks for Molecules
by Johannes Gasteiger, Florian Becker and Stephan Günnemann
Published at the Thirty-fifth Conference on Neural Information Processing Systems (NeurIPS), 2021

and

How Robust are Modern Graph Neural Network Potentials in Long and Hot Molecular Dynamics Simulations?
by Sina Stocker*, Johannes Gasteiger*, Florian Becker, Stephan Günnemann and Johannes T. Margraf
2022

* Both authors contributed equally to this research. Note that the author's name has changed from Johannes Klicpera to Johannes Gasteiger.

GemNet: Universal Directional Graph Neural Networks for Molecules

by Johannes Gasteiger, Florian Becker and Stephan Günnemann
Published at the Thirty-fifth Conference on Neural Information Processing Systems (NeurIPS), 2021

Abstract

Effectively predicting molecular interactions has the potential to accelerate molecular dynamics by multiple orders of magnitude and thus revolutionize chemical simulations. Graph neural networks (GNNs) have recently shown great successes for this task, overtaking classical methods based on fixed molecular kernels. However, they still appear very limited from a theoretical perspective, since regular GNNs cannot distinguish certain types of graphs. In this work we close this gap between theory and practice. We show that GNNs with directed edge embeddings and two-hop message passing are indeed universal approximators for predictions that are invariant to translation, and equivariant to permutation and rotation. We then leverage these insights and multiple structural improvements to propose the geometric message passing neural network (GemNet). We demonstrate the benefits of the proposed changes in multiple ablation studies. GemNet outperforms previous models on the COLL, MD17, and OC20 datasets by 34%, 41%, and 20%, respectively, and performs especially well on the most challenging molecules. Our implementation is available online.

Links

[Paper | GitHub (PyTorch) | GitHub (TensorFlow)]

Cite

Please cite our paper if you use the model, experimental results, or our code in your own work:

@inproceedings{gasteiger_gemnet_2021,
title = {GemNet: Universal Directional Graph Neural Networks for Molecules},
author = {Gasteiger, Johannes and Becker, Florian and G{\"u}nnemann, Stephan},
booktitle={Conference on Neural Information Processing Systems (NeurIPS)},
year = {2021}
}

How Robust are Modern Graph Neural Network Potentials in Long and Hot Molecular Dynamics Simulations?

by Sina Stocker*, Johannes Gasteiger*, Florian Becker, Stephan Günnemann and Johannes T. Margraf
2022

* Both authors contributed equally to this research.

Abstract

Graph neural networks (GNNs) have emerged as a powerful machine learning approach for the prediction of molecular properties. In particular, recently proposed advanced GNN models promise quantum chemical accuracy at a fraction of the computational cost. While the capabilities of such advanced GNNs have been extensively demonstrated on benchmark datasets, there have been few applications in real atomistic simulations. Here, we therefore put the robustness of GNN interatomic potentials to the test, using the recently proposed GemNet architecture as a testbed. Models are trained on the QM7-x database of organic molecules and used to perform extensive MD simulations. We find that low test set errors are not sufficient for obtaining stable dynamics and that severe pathologies sometimes only become apparent after hundreds of ps of dynamics. Nonetheless, highly stable and transferable GemNet potentials can be obtained with sufficiently large training sets.

Links

[Paper | GitHub (PyTorch) | GitHub (TensorFlow)]

Cite

Please cite our paper if you use the model, experimental results, or our code in your own work:

@article{stocker_gnn_2022,
title = {How Robust are Modern Graph Neural Network Potentials in Long and Hot Molecular Dynamics Simulations?},
author = {Stocker, Sina and Gasteiger, Johannes and Becker, Florian and G{\"u}nnemann, Stephan and Margraf, Johannes T.},
year = {2022}
}