Why Physics Still Matters: Improving Machine Learning Prediction of Material Properties with Phonon-Informed Datasets

2025-11-19

AI Summary

Material machine learning models suffer from limited predictive accuracy due to insufficient representation of low-symmetry atomic configurations—such as those induced by thermal excitation, defects, and chemical disorder—in training datasets. To address this, we propose a physics-guided data generation strategy grounded in phonon spectra: lattice vibrational modes inform biased sampling for graph neural networks (GNNs), enabling construction of a compact yet physically rich dataset of atomic configurations. Compared to large-scale random datasets, our approach significantly improves prediction accuracy for electronic structure and mechanical properties of optoelectronic materials at finite temperatures, while enhancing model interpretability by explicitly identifying critical bonding interactions. The core contribution is the establishment of a “physics-driven small-data” paradigm—demonstrating that targeted, physically informed datasets outperform unstructured large ones. This framework provides a generalizable, high-fidelity data curation methodology for material property prediction in energy conversion and photonics applications.

Technology Category

• Machine Learning: Graph-based Machine Learning
• Natural Language Processing: Generation
• Data Mining & Knowledge Management: Graph Mining, Social Network Analysis & Community

Application Category

• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Semantics and Knowledge: Methods, algorithms and applications for the development of semantic models, knowledge graphs and other forms of structured data models with machine-interpretable semantics
• Web Mining and Content Analysis: Large pretrained models with web data

Abstract

Machine learning (ML) methods have become powerful tools for predicting material properties with near first-principles accuracy and vastly reduced computational cost. However, the performance of ML models critically depends on the quality, size, and diversity of the training dataset. In materials science, this dependence is particularly important for learning from low-symmetry atomistic configurations that capture thermal excitations, structural defects, and chemical disorder, features that are ubiquitous in real materials but underrepresented in most datasets. The absence of systematic strategies for generating representative training data may therefore limit the predictive power of ML models in technologically critical fields such as energy conversion and photonics. In this work, we assess the effectiveness of graph neural network (GNN) models trained on two fundamentally different types of datasets: one composed of randomly generated atomic configurations and another constructed using physically informed sampling based on lattice vibrations. As a case study, we address the challenging task of predicting electronic and mechanical properties of a prototypical family of optoelectronic materials under realistic finite-temperature conditions. We find that the phonons-informed model consistently outperforms the randomly trained counterpart, despite relying on fewer data points. Explainability analyses further reveal that high-performing models assign greater weight to chemically meaningful bonds that control property variations, underscoring the importance of physically guided data generation. Overall, this work demonstrates that larger datasets do not necessarily yield better GNN predictive models and introduces a simple and general strategy for efficiently constructing high-quality training data in materials informatics.

Problem

• Improving ML prediction of material properties using phonon-informed datasets
• Addressing limitations in training data quality for low-symmetry atomistic configurations
• Developing efficient strategies for constructing high-quality training data in materials informatics

Innovation

• Uses phonon-informed sampling for dataset creation
• Applies graph neural networks to material property prediction
• Prioritizes chemically meaningful bonds through explainability analysis