Composable and adaptive design of machine learning interatomic potentials guided by Fisher-information analysis

2025-04-27

AI Summary

Addressing the challenge of balancing scalability, adaptability, and physical consistency in machine-learned interatomic potentials (MLIPs) for structurally diverse data, this work proposes a composable and adaptive modeling framework. The method employs iterative reconstruction of single-component models under unified physical constraints, integrated with a Fisher information matrix (FIM)-driven model evaluation and reconstruction guidance mechanism, alongside joint optimization of multi-property prediction errors. Notably, this is the first work to leverage the FIM for MLIP architecture selection and update—enabling modular extensibility and automatic hyperparameter adaptation. On a niobium multi-configuration dataset, the model achieves state-of-the-art accuracy with only 75 parameters: force RMSE = 0.172 eV/Å and energy RMSE = 0.013 eV/atom. The approach significantly enhances model flexibility, generalizability across configurations, and physical fidelity.

Technology Category

• Machine Learning: Learning with Manifolds
• Cognitive Modeling & Cognitive Systems: Adaptive Behavior
• Constraint Satisfaction and Optimization: Constraint Learning and Acquisition

Application Category

• User Modeling, Personalization and Recommendation: ML for personalized search and recommendations
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications

Abstract

An adaptive physics-informed model design strategy for machine-learning interatomic potentials (MLIPs) is proposed. This strategy follows an iterative reconfiguration of composite models from single-term models, followed by a unified training procedure. A model evaluation method based on the Fisher information matrix (FIM) and multiple-property error metrics is proposed to guide model reconfiguration and hyperparameter optimization. Combining the model reconfiguration and the model evaluation subroutines, we provide an adaptive MLIP design strategy that balances flexibility and extensibility. In a case study of designing models against a structurally diverse niobium dataset, we managed to obtain an optimal configuration with 75 parameters generated by our framework that achieved a force RMSE of 0.172 eV/{AA} and an energy RMSE of 0.013 eV/atom.

Problem

• Adaptive design of machine learning interatomic potentials
• Model evaluation using Fisher information and error metrics
• Optimizing MLIPs for diverse structural datasets

Innovation

• Adaptive physics-informed model design strategy
• Fisher information matrix guides optimization
• Iterative reconfiguration of composite models