Neural Structure Fields with Application to Crystal Structure Autoencoders

TL;DR

This paper introduces Neural Structure Fields (NeSF), a neural network-based continuous representation of crystal structures that improves accuracy and efficiency over traditional grid-based methods, enabling better inverse design of materials.

Contribution

The paper presents NeSF, a novel neural network approach that models crystal structures as continuous fields, overcoming resolution-complexity tradeoffs and enabling accurate structure reconstruction.

Findings

NeSF outperforms existing grid-based methods in accuracy.

NeSF can represent diverse crystal structures like perovskites and cuprates.

The autoencoder effectively reconstructs various crystal structures.

Abstract

Representing crystal structures of materials to facilitate determining them via neural networks is crucial for enabling machine-learning applications involving crystal structure estimation. Among these applications, the inverse design of materials can contribute to explore materials with desired properties without relying on luck or serendipity. We propose neural structure fields (NeSF) as an accurate and practical approach for representing crystal structures using neural networks. Inspired by the concepts of vector fields in physics and implicit neural representations in computer vision, the proposed NeSF considers a crystal structure as a continuous field rather than as a discrete set of atoms. Unlike existing grid-based discretized spatial representations, the NeSF overcomes the tradeoff between spatial resolution and computational complexity and can represent any crystal structure. We propose an autoencoder of crystal structures that can recover various crystal structures, such as those of perovskite structure materials and cuprate superconductors. Extensive quantitative results demonstrate the superior performance of the NeSF compared with the existing grid-based approach.