We introduce CrystalFormer, a transformer-based autoregressive model  specifically designed for space group-controlled generation of  crystalline materials. The space group symmetry significantly simplifies the crystal space, which is crucial for data and compute efficient  generative modeling of crystalline materials. Leveraging the prominent  discrete and sequential nature of the Wyckoff positions, CrystalFormer  learns to generate crystals by directly predicting the species and  locations of symmetry-inequivalent atoms in the unit cell. Our results  demonstrate that CrystalFormer matches state-of-the-art performance on  standard benchmarks for both validity, novelty, and stability of the  generated crystalline materials. Our analysis also shows that  CrystalFormer ingests sensible solid-state chemistry information from  data for generative modeling. The CrystalFormer unifies symmetry-based  structure search and generative pre-training in the realm of crystalline materials. The simplicity, generality, and flexibility of CrystalFormer position it as a promising architecture to be the foundational model of the entire crystalline materials space, heralding a new era in  materials modeling and discovery.

Please contact phweb@ust.hk should you have questions about the talk.