Crystal Diffusion Variational Autoencoder for Periodic Material Generation

Abstract

Generating the periodic structure of stable materials is a long-standing challenge for the material design community. This task is difficult because stable materials only exist in a low-dimensional subspace of all possible periodic arrangements of atoms: 1) the coordinates must lie in the local energy minimum defined by quantum mechanics, and 2) global stability also requires the structure to follow the complex, yet specific bonding preferences between different atom types. Existing methods fail to incorporate these factors and often lack proper invariances. We propose a Crystal Diffusion Variational Autoencoder (CDVAE) that captures the physical inductive bias of material stability. By learning from the data distribution of stable materials, the decoder generates materials in a diffusion process that moves atomic coordinates towards a lower energy state and updates atom types to satisfy bonding preferences between neighbors. Our model also explicitly encodes interactions across periodic boundaries and respects permutation, translation, rotation, and periodic invariances. We significantly outperform past methods in three tasks: 1) reconstructing the input structure, 2) generating valid, diverse, and realistic materials, and 3) generating materials that optimize a specific property. We also provide several standard datasets and evaluation metrics for the broader machine learning community.

Accepted to ICLR 2022. Code and data are publicly available at https://github.com/txie-93/cdvae

References in corpus (9)

• Diffusion Models Beat GANs on Image Synthesis
• A General-Purpose Machine Learning Framework for Predicting Properties of Inorganic Materials
• High-pressure phases of silane
• Inverse design of 3d molecular structures with conditional generative neural networks
• High-throughput discovery of novel cubic crystal materials using deep generative neural networks
• Rotation Invariant Graph Neural Networks using Spin Convolutions
• Crystal structure prediction of materials with high symmetry using differential evolution
• Study of Deep Generative Models for Inorganic Chemical Compositions
• E(n) Equivariant Normalizing Flows

Cited by in corpus (11)

• SELFIES and the future of molecular string representations
• Data-driven discovery of novel 2D materials by deep generative models
• Generative Language Model for Catalyst Discovery
• Recent progress in the JARVIS infrastructure for next-generation data-driven materials design
• Artificial Intelligence in Materials Science and Engineering: Current Landscape, Key Challenges, and Future Trajectorie
• Self-Supervised Generative Models for Crystal Structures
• Generative Design of Crystal Structures by Point Cloud Representations and Diffusion Model
• MC3D: The Materials Cloud computational database of experimentally known stoichiometric inorganics
• Zero Shot Molecular Generation via Similarity Kernels
• Discovery of 2D Materials via Symmetry-Constrained Diffusion Model
• From Polyhedra to Crystals: A Graph-Theoretic Framework for Crystal Structure Generation