Data-Driven Approach to Encoding and Decoding 3-D Crystal Structures

Abstract

Generative models have achieved impressive results in many domains including image and text generation. In the natural sciences, generative models have led to rapid progress in automated drug discovery. Many of the current methods focus on either 1-D or 2-D representations of typically small, drug-like molecules. However, many molecules require 3-D descriptors and exceed the chemical complexity of commonly used dataset. We present a method to encode and decode the position of atoms in 3-D molecules from a dataset of nearly 50,000 stable crystal unit cells that vary from containing 1 to over 100 atoms. We construct a smooth and continuous 3-D density representation of each crystal based on the positions of different atoms. Two different neural networks were trained on a dataset of over 120,000 three-dimensional samples of single and repeating crystal structures, made by rotating the single unit cells. The first, an Encoder-Decoder pair, constructs a compressed latent space representation of each molecule and then decodes this description into an accurate reconstruction of the input. The second network segments the resulting output into atoms and assigns each atom an atomic number. By generating compressed, continuous latent spaces representations of molecules we are able to decode random samples, interpolate between two molecules, and alter known molecules.

References in corpus (19)

• Auto-Encoding Variational Bayes
• Attention U-Net: Learning Where to Look for the Pancreas
• Large Scale GAN Training for High Fidelity Natural Image Synthesis
• Crystal Graph Convolutional Neural Networks for an Accurate and Interpretable Prediction of Material Properties
• SchNet - a deep learning architecture for molecules and materials
• Graph Networks as a Universal Machine Learning Framework for Molecules and Crystals
• Learning a Probabilistic Latent Space of Object Shapes via 3D Generative-Adversarial Modeling
• Junction Tree Variational Autoencoder for Molecular Graph Generation
• Learning to Discover Cross-Domain Relations with Generative Adversarial Networks
• Disentangling by Factorising
• Learning Representations and Generative Models for 3D Point Clouds
• Generative and Discriminative Voxel Modeling with Convolutional Neural Networks
• ChemTS: An Efficient Python Library for de novo Molecular Generation
• 3D Steerable CNNs: Learning Rotationally Equivariant Features in Volumetric Data
• Generative Models for Automatic Chemical Design
• Symmetry-adapted generation of 3d point sets for the targeted discovery of molecules
• PointFlow: 3D Point Cloud Generation with Continuous Normalizing Flows
• DEFactor: Differentiable Edge Factorization-based Probabilistic Graph Generation
• CrystalGAN: Learning to Discover Crystallographic Structures with Generative Adversarial Networks

Cited by in corpus (11)

• Generative adversarial networks (GAN) based efficient sampling of chemical space for inverse design of inorganic materials
• Constrained crystals deep convolutional generative adversarial network for the inverse design of crystal structures
• Machine Learning on Neutron and X-Ray Scattering
• Crystal Diffusion Variational Autoencoder for Periodic Material Generation
• Deep learning generative model for crystal structure prediction
• Real-time 3D Nanoscale Coherent Imaging via Physics-aware Deep Learning
• Crystal structure prediction of materials with high symmetry using differential evolution
• Study of Deep Generative Models for Inorganic Chemical Compositions
• On the Morality of Artificial Intelligence
• Predicting Material Properties Using a 3D Graph Neural Network with Invariant Local Descriptors
• Crystal structure prediction using age-fitness multi-objective genetic algorithm and coordination number constraints