Researchers have introduced MolCryst-MLIPs, an open database of machine-learned interatomic potentials (MLIPs) for molecular crystals. The initial release features fine-tuned MACE models for nine specific molecular crystal systems, developed using an automated machine learning pipeline. These models demonstrate high accuracy in predicting energy and forces, outperforming other foundation models in resolving polymorphic energy landscapes and maintaining structural integrity during simulations. AI
IMPACT Provides researchers with validated MLIPs for molecular crystal simulations, potentially accelerating materials discovery.
RANK_REASON The cluster contains an arXiv paper detailing a new database and models for molecular crystals. [lever_c_demoted from research: ic=1 ai=1.0]
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
