Tag: delta-learning

We introduced hierarchical machine learning (hML) approach for building highly accurate potential energy surfaces from multiple Δ-ML models, each trained on semi-automatically defined training points.