Crystalline materials often exhibit a high level of symmetry. However, most generative models do not account for symmetry, but rather model each atom without any constraints on its position or element. We propose a generative model, Wyckoff Diffusion (WyckoffDiff), which generates symmetry-based descriptions of crystals. This is enabled by considering a crystal structure representation that encodes all symmetry, and we design a novel neural network architecture which enables using this representation inside a discrete generative model framework. In addition to respecting symmetry by construction, the discrete nature of our model enables fast generation. We additionally present a new metric, Fréchet Wrenformer Distance, which captures the symmetry aspects of the materials generated, and we benchmark WyckoffDiff against recently proposed generative models for crystal generation. As a proof-of-concept study, we use WyckoffDiff to find new materials below the convex hull of thermodynamical stability.

Just as chatbots can generate text, artificial intelligence (AI) can also be used for generating new materials, a hot topic for, e.g., enabling new technology. In this work, we developed an AI model that can generate materials, ensuring that the generated materials have the characteristic symmetrical properties that are found in materials in nature.Instead of viewing materials as atoms in space which in principle can be placed anywhere, we used a description of materials that explicitly includes information about symmetry in materials, so that the model does not have to learn this property by itself. As a proof of concept, we used our materials in a materials generation workflow, where the materials proposed by our model were further validated, and we find several new materials which are also predicted to be stable, i.e., they could exist in nature.This model is a new direction in generating materials, and while it can serve as inspiration for future work, it has already shown potential in materials discovery.