SymmCD: Symmetry-Preserving Crystal Generation with Diffusion Models

• DOI: 10.48550/arxiv.2502.03638
• Publication Date: 2025-02-05
• Authors: Daniel Lévy; Siba Smarak Panig...; Sékou-Oumar Kaba; Qiang Zhu; Kin Long Kelvin Lee; Mikhail Galkin; Santiago Miret; Siamak Ravanbakhsh
• Keywords: Crystal (programming language)

Generating novel crystalline materials has potential to lead advancements in fields such as electronics, energy storage, and catalysis. The defining characteristic of crystals is their symmetry, which plays a central role determining physical properties. However, existing crystal generation methods either fail generate that display the symmetries real-world crystals, or simply replicate symmetry information from examples database. To address this limitation, we propose SymmCD, diffusion-based generative model explicitly incorporates crystallographic into process. We decompose two components learn joint distribution through diffusion: 1) asymmetric unit, smallest subset can whole transformations, and; 2) transformations needed be applied each atom unit. also use interpretable representation for these enabling generalization across different groups. showcase competitive performance SymmCD on Materials Project, obtaining diverse valid with realistic predicted