• Corpus ID: 9665943

Neural Message Passing for Quantum Chemistry

  title={Neural Message Passing for Quantum Chemistry},
  author={Justin Gilmer and Samuel S. Schoenholz and Patrick F. Riley and Oriol Vinyals and George E. Dahl},
Supervised learning on molecules has incredible potential to be useful in chemistry, drug discovery, and materials science. [] Key Method In this paper, we reformulate existing models into a single common framework we call Message Passing Neural Networks (MPNNs) and explore additional novel variations within this framework. Using MPNNs we demonstrate state of the art results on an important molecular property prediction benchmark; these results are strong enough that we believe future work should focus on…

Figures and Tables from this paper

Flexible dual-branched message passing neural network for quantum mechanical property prediction with molecular conformation

The proposed dualbranched neural network for molecular property prediction based on message-passing framework outperforms other recent models with sparser representations and indicates that in the chemical property prediction tasks, the diverse chemical nature of targets should be carefully considered for both model performance and generalizability.

Message-passing neural networks for high-throughput polymer screening

It is shown that message-passing neural networks trained with and without 3D structural information for these molecules achieve similar accuracy, comparable to state-of-the-art methods on existing benchmark datasets, and learned molecular representations can be leveraged to reduce the training data required to transfer predictions to a new density functional theory functional.

Flexible Dual-Branched Message-Passing Neural Network for a Molecular Property Prediction

This study proposes a dual-branched neural network for molecular property prediction based on both the message-passing framework and standard multilayer perceptron neural networks, and introduces a discrete branch to learn single-atom features without local aggregation.

GemNet: Universal Directional Graph Neural Networks for Molecules

This work shows 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, and proposes the geometric message passing neural network (GemNet).

Gated Graph Recursive Neural Networks for Molecular Property Prediction

This work proposes a simple and powerful graph neural networks for molecular property prediction as a directed complete graph in which each atom has a spatial position, and introduces a recursive neural network with simple gating function.

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

This work proposes the polarizable atom interaction neural network (PAINN) and improves on common molecule benchmarks over previous networks, while reducing model size and inference time and leverage the equivariant atomwise representations obtained by PAINN for the prediction of tensorial properties.

Message passing neural networks for molecular property prediction

This thesis applies the Direct Message Passing Neural Network (D-MPNN) from [47, 48] to 19 publicly available property prediction datasets, and it is demonstrated that it consistently outperforms prior machine learning models.

Are Learned Molecular Representations Ready For Prime Time?

A graph convolutional model is introduced that consistently matches or outperforms models using fixed molecular descriptors as well as previous graph neural architectures on both public and proprietary datasets.



Quantum-chemical insights from deep tensor neural networks

An efficient deep learning approach is developed that enables spatially and chemically resolved insights into quantum-mechanical observables of molecular systems, and unifies concepts from many-body Hamiltonians with purpose-designed deep tensor neural networks, which leads to size-extensive and uniformly accurate chemical space predictions.

Deep Architectures and Deep Learning in Chemoinformatics: The Prediction of Aqueous Solubility for Drug-Like Molecules

A brief overview of deep learning methods is presented and in particular how recursive neural network approaches can be applied to the problem of predicting molecular properties, by considering an ensemble of recursive neural networks associated with all possible vertex-centered acyclic orientations of the molecular graph.

Learning Invariant Representations of Molecules for Atomization Energy Prediction

This paper adopts a learning-from-scratch approach where quantum-mechanical molecular energies are predicted directly from the raw molecular geometry, and suggests a benefit from setting flexible priors and enforcing invariance stochastically rather than structurally.

Machine Learning Predictions of Molecular Properties: Accurate Many-Body Potentials and Nonlocality in Chemical Space

A systematic hierarchy of efficient empirical methods to estimate atomization and total energies of molecules and is achieved by a vectorized representation of molecules (so-called Bag of Bonds model) that exhibits strong nonlocality in chemical space.

Molecular graph convolutions: moving beyond fingerprints

Molecular graph convolutions are described, a machine learning architecture for learning from undirected graphs, specifically small molecules, that represent a new paradigm in ligand-based virtual screening with exciting opportunities for future improvement.

Communication: Understanding molecular representations in machine learning: The role of uniqueness and target similarity.

This work introduces a hierarchy of representations which meet uniqueness and target similarity criteria, and relies on interatomic many body expansions, as implemented in universal force-fields, including Bonding, Angular (BA), and higher order terms to systematically control target similarity.

Gated Graph Sequence Neural Networks

This work studies feature learning techniques for graph-structured inputs and achieves state-of-the-art performance on a problem from program verification, in which subgraphs need to be matched to abstract data structures.

Automatic Generation of Complementary Descriptors with Molecular Graph Networks

We describe a method for the automatic generation of weakly correlated descriptors for molecular data sets. The method can be regarded as a statistical learning procedure that turns the molecular

The Graph Neural Network Model

A new neural network model, called graph neural network (GNN) model, that extends existing neural network methods for processing the data represented in graph domains, and implements a function tau(G,n) isin IRm that maps a graph G and one of its nodes n into an m-dimensional Euclidean space.

Fast and accurate modeling of molecular atomization energies with machine learning.

A machine learning model is introduced to predict atomization energies of a diverse set of organic molecules, based on nuclear charges and atomic positions only, and applicability is demonstrated for the prediction of molecular atomization potential energy curves.