Back to papers
October 26, 2026cs.LGcs.AI

Multi-Fidelity Training Reshapes Uncertainty Decomposition in Atomistic Models

Categories

cs.LG, cs.AI