Machine-learned interatomic potential

Sat, 18 Jul 2026 10:06:25 +0000

Definition

These potentials enable molecular dynamics simulations at near-quantum accuracy but with classical computational speed. By training on high-fidelity data from density functional theory (DFT), they allow researchers to simulate larger systems over longer timescales. This is crucial for materials science, chemistry, and biology, facilitating the discovery of new materials and understanding complex molecular interactions that were previously computationally prohibitive to model accurately.

Summary

An ML-based mathematical function that predicts the forces and energies between atoms, serving as a surrogate for expensive quantum mechanical calculations.

Science on English AI Terms Dictionary

Machine-learned interatomic potential

Definition

Summary