New OPAL-Bench benchmark offers comprehensive assessment of crystal structure prediction methods

Researchers have introduced OPAL-Bench, a new benchmark designed to quantitatively assess Crystal Structure Prediction (CSP) methods. This benchmark addresses limitations in existing frameworks by incorporating threshold-free geometric and physical distance metrics, and by accounting for polymorphism. The study found that no single method excelled across all metrics, and geometric performance did not strongly correlate with physical plausibility. Additionally, the paper introduces OPAL-CSP, an SE(3)-equivariant flow-matching baseline that shows competitive performance on geometric and physical metrics. AI

IMPACT Provides a more robust framework for evaluating and advancing generative models in materials science.

RANK_REASON The cluster contains an academic paper detailing a new benchmark and baseline method for crystal structure prediction. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.AI →

• Gabriel González
• OPAL-Bench
• OPAL-CSP
• SE(3)-Transformers: 3D Roto-Translation Equivariant Attention Networks

• paper
• other

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →