PFT: Phonon Fine-tuning for Machine Learned Interatomic Potentials
This addresses the issue of inaccurate higher-order derivative predictions in materials science, offering a domain-specific incremental improvement for computational materials modeling.
The paper tackled the problem of machine learned interatomic potentials (MLIPs) having errors in curvature that degrade vibrational property predictions by introducing phonon fine-tuning (PFT), which directly supervises second-order force constants, resulting in a 55% average improvement on phonon thermodynamic properties and state-of-the-art performance on the MDR Phonon benchmark.
Many materials properties depend on higher-order derivatives of the potential energy surface, yet machine learned interatomic potentials (MLIPs) trained with standard a standard loss on energy, force, and stress errors can exhibit error in curvature, degrading the prediction of vibrational properties. We introduce phonon fine-tuning (PFT), which directly supervises second-order force constants of materials by matching MLIP energy Hessians to DFT-computed force constants from finite displacement phonon calculations. To scale to large supercells, PFT stochastically samples Hessian columns and computes the loss with a single Hessian-vector product. We also use a simple co-training scheme to incorporate upstream data to mitigate catastrophic forgetting. On the MDR Phonon benchmark, PFT improves Nequix MP (trained on Materials Project) by 55% on average across phonon thermodynamic properties and achieves state-of-the-art performance among models trained on Materials Project trajectories. PFT also generalizes to improve properties beyond second-derivatives, improving thermal conductivity predictions that rely on third-order derivatives of the potential energy.