SOURCE-LINKED INTELLIGENCE

A budget-dependent crossover between coverage- and response-based training-set selection for machine-learned interatomic potentials

arXiv · AI, language, vision and robotics · article · Sep 5, 2026 · UTC

Selecting compact training sets for machine-learned interatomic potentials requires deciding whether to preserve structural diversity or target configurations on which models disagree. The better choice can depend on how much data is retained, making a comparison at one training-set size insufficient. Here we link selection criteria to prediction accuracy through a budget-resolved comparison of retrained MACE models on GAP-20 Carbon and pooled revised MD17. Structural coverage is compared with a response-guided selector that targets disagreement between a coverage-trained model and a full-data

Read original source ↗ Open in workspace

recordType

paper

region

Global