Overcoming systematic softening in universal machine learning interatomic potentials by fine-tuning

TL;DR

This work reveals a universal PES softening across three foundational uMLIPs (M3GNet, CHGNet, MACE-MP-0), where energies and forces are systematically underpredicted in high-energy, out-of-distribution environments due to biased pretraining data. The authors provide a mechanistic explanation linking softening to limited PES sampling in the Materials Project dataset and demonstrate that a minimal fine-tuning protocol—specifically, a simple linear correction using a single high-energy data point—substantially mitigates the issue across surfaces, defects, solid-solution energetics, phonons, and ion migration barriers. They quantify the softening with a scale parameter and show that most chemistries exhibit $s<1$, confirming a broad, systematic bias. The findings offer a practical, data-efficient route to correct uMLIPs and argue for richer PES-sampling in future foundational datasets, with significant implications for reliable, scalable atomistic simulations and materials discovery.

Abstract

Machine learning interatomic potentials (MLIPs) have introduced a new paradigm for atomic simulations. Recent advancements have seen the emergence of universal MLIPs (uMLIPs) that are pre-trained on diverse materials datasets, providing opportunities for both ready-to-use universal force fields and robust foundations for downstream machine learning refinements. However, their performance in extrapolating to out-of-distribution complex atomic environments remains unclear. In this study, we highlight a consistent potential energy surface (PES) softening effect in three uMLIPs: M3GNet, CHGNet, and MACE-MP-0, which is characterized by energy and force under-prediction in a series of atomic-modeling benchmarks including surfaces, defects, solid-solution energetics, phonon vibration modes, ion migration barriers, and general high-energy states. We find that the PES softening behavior originates from a systematic underprediction error of the PES curvature, which derives from the biased sampling of near-equilibrium atomic arrangements in uMLIP pre-training datasets. We demonstrate that the PES softening issue can be effectively rectified by fine-tuning with a single additional data point. Our findings suggest that a considerable fraction of uMLIP errors are highly systematic, and can therefore be efficiently corrected. This result rationalizes the data-efficient fine-tuning performance boost commonly observed with foundational MLIPs. We argue for the importance of a comprehensive materials dataset with improved PES sampling for next-generation foundational MLIPs.

Figures (6)

• Figure 1: Potential energy surface softening in uMLIPs. Left: schematic representation of the potential energy surface (PES) described in density functional theory (DFT), with two arbitrary coordinate axes. Right: PES described by universal machine learning interatomic potentials (uMLIPs), which well describes the PES regions sampled by near-equilibrium states in the pre-training dataset (orange), but experience larger errors in high-energy regions (red) with under-prediction of energies and forces. The softening behavior is largely systematic and, therefore can be efficiently fixed by a linear correction with a small amount of data augmentation.
• Figure 2: uMLIP performance on surfaces, defects, and solid solutions. (a) Comparison of DFT surface energies and MLIP surface energies, evaluated on 147 surfaces from 29 chemical systems. (b) Comparison of DFT defect energies and MLIP defect energies, evaluated on 134 point defects from 32 chemical systems. (c) Formation energies in Ca$_x$Mg$_{2-x}$O$_2$ solid solution from DFT and uMLIPs. Each point corresponds to the energy of a specific Ca-Mg cation arrangement at a given Ca fraction. The distributions of energies are collectively underestimated, which would lead to an underprediction of the miscibility gap temperature in uMLIPs compared to DFT.
• Figure 3: Softened phonon vibration modes in uMLIPs. (a) the phonon dispersion relation and density of states (DOS) of Cs$_4$F$_4$(mp-1784) calculated with DFT and uMLIPs. Systematic underpredictions of phonon vibration frequencies are observed with all uMLIPs. (b) Distribution of ratios between uMLIP maximum frequency to DFT maximum frequency for 229 different compounds.
• Figure 4: Underpredicted ion migration barriers in DFT and uMLIPs. (a) An example of a Mg-ion migration path in V2O3(SO4)2 (mp-28207) with 6 intermediate images. The migration barrier is defined as the energy difference between the highest and lowest image. (b) The distribution of 477 energy barrier differences between uMLIPs and DFT, showing uMLIPs' tendency to underestimate the ion migration barriers.
• Figure 5: The PES softening scale from shifted force predictions. (a) uMLIP forces vs. DFT forces in high-energy states, sampled from high-temperature MDs of $\ch{Li2B3PO8}$(mp-1020015). Systematic softening of PES is indicated by the tilted distribution of forces from the diagonal. The softening scale is defined as the slope of the distribution, where softening is indicated by slope $<$ 1 . cMAE stands for corrected mean absolute error, which is the MAE if the softening scale is corrected to 1, equivalent to having the force distribution rotated back to diagonal. (b) Distribution of softening scales of 1000 compounds sampled from the WBM dataset, showing the PES softening behavior is universal across various chemical systems.