AI Paper Atlas

Research map / Architectures and efficiency / Pruning and efficient tuning

PFT: Phonon Fine-tuning for Machine Learned Interatomic Potentials

Teddy Koker, Abhijeet Sadashiv Gangan, Mit Kotak, Jaime Marian, Tess Smidt

ICML 2026 · Architectures and efficiency › Pruning and efficient tuning · approach: Adapter-based parameter-efficient tuning · 0 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
fine-tune with Hessian (force constant) loss PFTNequix MP · MDR Phonon testaverage MAE across phonon properties: 55% average reduction; max frequency 24 -> 12 KImproved
apply PFT to a stronger base trained on OMat24 data (Nequix OAM)Nequix OAM · MDR Phonon testaverage MAE across phonon properties: 50% average relative improvement; max frequency 17 -> 9 KImproved
fine-tune energy/force/stress directly on phonon displacement data instead of Hessian loss (ablation)Nequix MP · MDR Phonon testmax phonon frequency MAE (K): 24 -> 182Hurt
add co-training with upstream MPtrj data (ablation)Nequix MP PFT · Matbench Discovery stabilityF1: 0.301 -> 0.748Improved
add co-training with upstream MPtrj data (ablation)Nequix MP PFT · MDR Phonon testmax phonon frequency MAE (K): 10 -> 12Hurt
add co-training with upstream MPtrj data (ablation)Nequix MP PFT · Matbench Discovery geometry optimizationRMSD (A): 0.089 -> 0.087Improved
compute force constants analytically via autodiff instead of finite displacement (ablation)Nequix MP PFT · MDR Phonon testphonon property MAE: max frequency 12 -> 12No clear effect
apply PFT to MACE-MP-0 without co-trainingMACE-MP-0 · MDR Phonon testmax phonon frequency MAE (K): 61 -> 19Improved
apply PFT (generalization to third-order force constants)Nequix MP · third-order force constantsMAE of Phi(3) (meV/A^3): 10.52 -> 8.35Improved
apply PFT to improve thermal conductivityNequix MP · Matbench Discovery thermal conductivitykappa SRME: 0.446 -> 0.307Improved
apply PFT to improve thermal conductivityNequix OAM · Matbench Discovery thermal conductivitykappa SRME: 0.267 -> 0.198Improved
apply PFT with co-trainingNequix MP · Matbench Discovery stabilityF1: 0.751 -> 0.748No clear effect

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy