PFT: Phonon Fine-tuning for Machine Learned Interatomic Potentials

• URL: http://arxiv.org/abs/2601.07742v2
• Date: Thu, 15 Jan 2026 21:43:37 GMT
• Title: PFT: Phonon Fine-tuning for Machine Learned Interatomic Potentials
• Authors: Teddy Koker, Abhijeet Gangan, Mit Kotak, Jaime Marian, Tess Smidt,
• Abstract summary: We introduce phonon fine-tuning (PFT), which directly supervises second-order force constants of materials.<n>PFT generalizes to improve properties beyond second-derivatives, improving thermal conductivity predictions.
• Score: 1.9466051980292256
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: 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.

Related papers

• From Regression to Classification: Exploring the Benefits of Categorical Representations of Energy in MLIPs [1.0998907972211756]
  Density Functional Theory (DFT) is a widely used computational method for estimating the energy and behavior of molecules.<n>Machine Learning Interatomic Potentials (MLIPs) are models trained to approximate DFT-level energies and forces at dramatically lower computational cost.<n>In this work, we explore a multi-class classification formulation that predicts a categorical distribution over energy/force values.
  arXiv  Detail & Related papers  (2025-12-01T00:36:42Z)
• FIRE-GNN: Force-informed, Relaxed Equivariance Graph Neural Network for Rapid and Accurate Prediction of Surface Properties [8.537263229630897]
  We introduce FIRE-GNN, which integrates surface-normal symmetry breaking and machine learning interatomic potential (MLIP)-derived force information.<n>It achieves a twofold reduction in mean absolute error (down to 0.065 eV) over the previous state-of-the-art for work function prediction.<n>This model enables accurate and rapid predictions of the work function and cleavage energy across a vast chemical space.
  arXiv  Detail & Related papers  (2025-08-22T00:07:52Z)
• Constructing accurate machine-learned potentials and performing highly efficient atomistic simulations to predict structural and thermal properties [6.875235178607604]
  We introduce a neuroevolution potential (NEP) trained on a dataset generated from ab initio molecular dynamics (AIMD) simulations. We calculate the phonon density of states (DOS) and radial distribution function (RDF) using both machine learning potentials. While the MTP potential offers slightly higher accuracy, the NEP achieves a remarkable 41-fold increase in computational speed.
  arXiv  Detail & Related papers  (2024-11-16T23:16:59Z)
• Predicting ionic conductivity in solids from the machine-learned potential energy landscape [68.25662704255433]
  We propose an approach for the quick and reliable screening of ionic conductors through the analysis of a universal interatomic potential.<n>Eight out of the ten highest-ranked materials are confirmed to be superionic at room temperature in first-principles calculations.<n>Our method achieves a speed-up factor of approximately 50 compared to molecular dynamics driven by a machine-learning potential, and is at least 3,000 times faster compared to first-principles molecular dynamics.
  arXiv  Detail & Related papers  (2024-11-11T09:01:36Z)
• Non-asymptotic Convergence of Training Transformers for Next-token Prediction [48.9399496805422]
  Transformers have achieved extraordinary success in modern machine learning due to their excellent ability to handle sequential data. This paper provides a fine-grained non-asymptotic analysis of the training dynamics of a one-layer transformer. We show that the trained transformer presents non-token prediction ability with dataset shift.
  arXiv  Detail & Related papers  (2024-09-25T20:22:06Z)
• Higher-Order Equivariant Neural Networks for Charge Density Prediction in Materials [3.7655047338409893]
  ChargE3Net is an E(3)-equivariant graph neural network for predicting electron density in atomic systems. We show that ChargE3Net exceeds the performance of prior work on diverse sets of molecules and materials.
  arXiv  Detail & Related papers  (2023-12-08T21:56:19Z)
• SIP: Injecting a Structural Inductive Bias into a Seq2Seq Model by Simulation [75.14793516745374]
  We show how a structural inductive bias can be efficiently injected into a seq2seq model by pre-training it to simulate structural transformations on synthetic data. Our experiments show that our method imparts the desired inductive bias, resulting in better few-shot learning for FST-like tasks.
  arXiv  Detail & Related papers  (2023-10-01T21:19:12Z)
• FC-PINO: High Precision Physics-Informed Neural Operators via Fourier Continuation [60.706803227003995]
  We introduce the FC-PINO (Fourier-Continuation-based PINO) architecture which extends the accuracy and efficiency of PINO to non-periodic and non-smooth PDEs.<n>We demonstrate that standard PINO struggles to solve non-periodic and non-smooth PDEs with high precision, across challenging benchmarks.<n>In contrast, the proposed FC-PINO provides accurate, robust, and scalable solutions, substantially outperforming PINO alternatives.
  arXiv  Detail & Related papers  (2022-11-29T06:37:54Z)
• Learning Physics-Informed Neural Networks without Stacked Back-propagation [82.26566759276105]
  We develop a novel approach that can significantly accelerate the training of Physics-Informed Neural Networks. In particular, we parameterize the PDE solution by the Gaussian smoothed model and show that, derived from Stein's Identity, the second-order derivatives can be efficiently calculated without back-propagation. Experimental results show that our proposed method can achieve competitive error compared to standard PINN training but is two orders of magnitude faster.
  arXiv  Detail & Related papers  (2022-02-18T18:07:54Z)
• Comparison of Optical Response from DFT Random Phase Approximation and Low-Energy Effective Model: Strained Phosphorene [0.0]
  We compare and contrast the dispersive permittivity tensor, using both a low-energy effective model and density functional theory (DFT) Our results suggest that the random-phase approximation employed in widely used DFT packages should be revisited and improved to be able to predict these fundamental electronic characteristics of a given material with confidence.
  arXiv  Detail & Related papers  (2021-09-01T18:00:06Z)
• Exponentially Weighted l_2 Regularization Strategy in Constructing Reinforced Second-order Fuzzy Rule-based Model [72.57056258027336]
  In the conventional Takagi-Sugeno-Kang (TSK)-type fuzzy models, constant or linear functions are usually utilized as the consequent parts of the fuzzy rules. We introduce an exponential weight approach inspired by the weight function theory encountered in harmonic analysis.
  arXiv  Detail & Related papers  (2020-07-02T15:42:15Z)
• Training Deep Energy-Based Models with f-Divergence Minimization [113.97274898282343]
  Deep energy-based models (EBMs) are very flexible in distribution parametrization but computationally challenging. We propose a general variational framework termed f-EBM to train EBMs using any desired f-divergence. Experimental results demonstrate the superiority of f-EBM over contrastive divergence, as well as the benefits of training EBMs using f-divergences other than KL.
  arXiv  Detail & Related papers  (2020-03-06T23:11:13Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.