SA-GAT-SR: Self-Adaptable Graph Attention Networks with Symbolic Regression for high-fidelity material property prediction

2025-05-01

AI Summary

To address the challenge of balancing predictive accuracy and physical interpretability in high-throughput property prediction for materials, this paper proposes an end-to-end interpretable learning framework integrating a Self-Adaptive Graph Attention Network (SA-GAT) with Symbolic Regression (SR). The method introduces a novel adaptive feature screening and encoding algorithm that automatically identifies physically meaningful variables from an 180-dimensional high-dimensional descriptor space. It achieves, for the first time, joint optimization of GAT and SR to directly generate closed-form analytical expressions grounded in quantum-mechanical principles. Employing an O(n) lightweight architecture, the framework maintains high-fidelity prediction accuracy (MAE ≤ 0.08 eV/atom) while accelerating symbolic regression by 23×, substantially enhancing both physical insight and computational efficiency.

Technology Category

• Machine Learning: Neuro-Symbolic Learning
• Cognitive Modeling & Cognitive Systems: Symbolic Representations
• Knowledge Representation and Reasoning: Diagnosis and Abductive Reasoning

Application Category

• Search and Retrieval-Augmented AI: Web query analysis, representation and understanding
• Graph Algorithms and Modeling for the Web: Algorithms and analysis for heterogeneous, signed, attributed, multi-relational, temporal, higher-order, and annotated 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

Abstract

Recent advances in machine learning have demonstrated an enormous utility of deep learning approaches, particularly Graph Neural Networks (GNNs) for materials science. These methods have emerged as powerful tools for high-throughput prediction of material properties, offering a compelling enhancement and alternative to traditional first-principles calculations. While the community has predominantly focused on developing increasingly complex and universal models to enhance predictive accuracy, such approaches often lack physical interpretability and insights into materials behavior. Here, we introduce a novel computational paradigm, Self-Adaptable Graph Attention Networks integrated with Symbolic Regression (SA-GAT-SR), that synergistically combines the predictive capability of GNNs with the interpretative power of symbolic regression. Our framework employs a self-adaptable encoding algorithm that automatically identifies and adjust attention weights so as to screen critical features from an expansive 180-dimensional feature space while maintaining O(n) computational scaling. The integrated SR module subsequently distills these features into compact analytical expressions that explicitly reveal quantum-mechanically meaningful relationships, achieving 23 times acceleration compared to conventional SR implementations that heavily rely on first principle calculations-derived features as input. This work suggests a new framework in computational materials science, bridging the gap between predictive accuracy and physical interpretability, offering valuable physical insights into material behavior.

Problem

• Combines GNNs with symbolic regression for interpretable material property prediction
• Automatically identifies critical features from a high-dimensional space efficiently
• Bridges gap between predictive accuracy and physical interpretability in materials science

Innovation

• Self-Adaptable Graph Attention Networks adjust attention weights
• Symbolic Regression distills features into analytical expressions
• Combines GNNs with SR for interpretable material prediction