SymmBFN adapts Bayesian Flow Networks to generate crystals with explicit space-group constraints, reporting up to 50 times faster stable-structure generation than prior diffusion models.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Symmetry-Aware Bayesian Flow Networks for Crystal Generation
SymmBFN adapts Bayesian Flow Networks to generate crystals with explicit space-group constraints, reporting up to 50 times faster stable-structure generation than prior diffusion models.