Risk-Controlled DFT Selection for Machine-Learned Interatomic Potentials

Automated materials discovery requires evaluating large pools of candidate configurations, with density functional theory (DFT) as a major throughput bottleneck.