Robust and Interpretable Adaptation of Equivariant Materials Foundation Models via Sparsity-promoting Fine-tuning

2026-06-17

AI Summary

This work addresses the challenge of adapting pretrained foundation models for materials science across diverse domains, where differences in physicochemical properties and computational settings hinder direct transfer. The authors propose a sparsity-promoting fine-tuning approach that leverages the structural priors of E(3)-equivariant neural networks, updating only 0.5%–3% of the model parameters to achieve efficient domain adaptation. By integrating low-rank adaptation with physics-informed constraints, the method matches or exceeds the performance of full-parameter fine-tuning in predicting energies and forces for both molecular and crystalline systems. Furthermore, it generalizes successfully to magnetic moment prediction and magnetism-aware total energy modeling, demonstrating strong versatility and interpretability. Notably, the learned sparse adaptation patterns exhibit clear physical meaning, enhancing model transparency and scientific insight.

Abstract

Pre-trained materials foundation models, or machine learning interatomic potentials, leverage general physicochemical knowledge to effectively approximate potential energy surfaces. However, they often require domain-specific calibration due to physicochemical diversity as well as mismatches between practical computational settings and those used in constructing the pre-training data. To address this, we propose a sparsity-promoting fine-tuning method that selectively updates model parameters by exploiting the structural properties of E(3)-equivariant materials foundation models. On energy and force prediction tasks across molecular and crystalline benchmarks, our method matches or surpasses full fine-tuning and equivariant low-rank adaptation while updating only $\sim$3~\% of parameters, and in some cases as little as $\sim$0.5~\%. Beyond energy and force calibration, we further demonstrate task generalizability by applying our method to magnetic moment prediction and magnetism-aware total energy modeling. Finally, analysis of sparsity patterns reveals physically interpretable signatures, such as enhanced $d$-orbital contributions in transition metal systems. Overall, our results establish sparsity-promoting fine-tuning as a flexible and interpretable method for domain specialization of equivariant materials foundation models.

Problem

• materials foundation models
• domain adaptation
• equivariance
• fine-tuning
• sparsity

Innovation

• sparsity-promoting fine-tuning
• E(3)-equivariant models
• materials foundation models
• interpretable adaptation
• parameter-efficient transfer learning