MLIP Performance Analysis
Performance datasets and analysis notebooks for the publication “Learning from models: high-dimensional analyses on the performance of machine learning interatomic potentials.” The repository includes the performance datasets used in the paper along with notebooks reproducing each figure and analysis, covering methods such as Pareto front computation, inverted generational distance, and the Cholesky method.
Reference: Yunsheng Liu, Yifei Mo*, “Learning from models: high-dimensional analyses on the performance of machine learning interatomic potentials”, NPJ Computational Materials, 10, 159 (2024)