Researchers have developed a new generative model called Coarse-Grained Open Materials Generation (CG-OMatG) to tackle the challenge of predicting molecular crystal structures. This model utilizes an equivariant Riemannian flow-based approach with a hierarchical representation, treating molecules as rigid bodies and predicting their positions, orientations, and lattice parameters. Trained on the Open Molecular Crystals (OMC25) and Cambridge Structural Database (CSD), CG-OMatG is further refined using reinforcement learning to generate lower-energy structures. The model's performance was validated on a crystal structure prediction benchmark, showing promise for accelerating materials discovery. AI
IMPACT This research could accelerate the discovery of new organic solid-state materials by improving the efficiency of polymorph screening.
RANK_REASON The cluster contains an academic paper detailing a new model and methodology for a scientific problem. [lever_c_demoted from research: ic=1 ai=1.0]
- arXiv
- Cambridge Structural Database
- Coarse-Grained Open Materials Generation
- COMPACK
- OMC25
- Open Molecular Crystals
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
