Role of Structural and Conformational Diversity for Machine Learning Potentials
This work addresses data generation challenges for researchers in computational chemistry and materials science, offering incremental guidelines to improve model generalization.
The study investigated how structural and conformational diversity in training data affects the generalization of Machine Learning Interatomic Potentials, finding that a balance between these diversities is crucial for optimal performance, but current Quantum Mechanics datasets fail to achieve this trade-off.
In the field of Machine Learning Interatomic Potentials (MLIPs), understanding the intricate relationship between data biases, specifically conformational and structural diversity, and model generalization is critical in improving the quality of Quantum Mechanics (QM) data generation efforts. We investigate these dynamics through two distinct experiments: a fixed budget one, where the dataset size remains constant, and a fixed molecular set one, which focuses on fixed structural diversity while varying conformational diversity. Our results reveal nuanced patterns in generalization metrics. Notably, for optimal structural and conformational generalization, a careful balance between structural and conformational diversity is required, but existing QM datasets do not meet that trade-off. Additionally, our results highlight the limitation of the MLIP models at generalizing beyond their training distribution, emphasizing the importance of defining applicability domain during model deployment. These findings provide valuable insights and guidelines for QM data generation efforts.