Comparative Assessment of Composition-and Structure-Based Surrogate Models Across 2D Materials Databases

Key Points

• The research aims to evaluate the accuracy and transferability of composition-based and structure-based surrogate models in materials discovery.
• Benchmarking various surrogate models across different 2D materials databases.
• Comparative analysis of model performance based on accuracy metrics.
• Utilization of machine learning techniques to enhance model predictions.
• Composition-based models show high accuracy in certain datasets but struggle with transferability.
• Structure-based models maintain better transferability across heterogeneous databases.
• Overall performance varies significantly based on the model type and the materials database used.

Abstract