Equivariant message passing for the prediction of tensorial properties and molecular spectra
This work provides a more data-efficient and faster method for predicting chemical properties and simulating molecular spectra, which is significant for researchers in computational chemistry and materials science.
This paper addresses the data inefficiency of message passing neural networks (MPNNs) for chemical property prediction by extending the formulation to rotationally equivariant representations. The proposed Polarizable Atom Interaction Neural Network (PaiNN) improves on common molecule benchmarks, reduces model size and inference time, and achieves 4-5 orders of magnitude speedup in simulating molecular spectra compared to electronic structure references.
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.