Structured Extraction of Process Structure Properties Relationships in Materials Science

2025-04-04

AI Summary

Efficient extraction of process–structure–property (PSP) relationships from unstructured materials science literature remains challenging due to the lack of standardized annotation schemes and domain-adapted extraction frameworks. Method: This work introduces the first general-purpose annotation schema for PSP ternary relations and proposes a cross-domain structured knowledge extraction framework covering high-temperature materials and microstructural simulation uncertainty quantification. The method integrates MatBERT-CRF sequence labeling with fine-tuned GPT-4o, leveraging cross-domain mixed training and zero-shot transfer strategies. Results: On Domain I, the fine-tuned GPT-4o achieves 89.2% F1 for entity extraction—significantly outperforming MatBERT-CRF. Incorporating Domain II data narrows the performance gap, demonstrating both the annotation schema’s strong generalizability and the complementary strengths of the two approaches. This work establishes a scalable technical paradigm for materials knowledge graph construction and trustworthy large-language-model reasoning.

Technology Category

• Natural Language Processing: Information Extraction
• Data Mining & Knowledge Management: Linked Open Data, Knowledge Graphs & KB Completion
• Knowledge Representation and Reasoning: Geometric, Spatial, and Temporal Reasoning

Application Category

• Semantics and Knowledge: Scalable techniques for the creation, curation, publication, maintenance, and consumption of large, Web-based, structured, reusable, knowledge graphs and ontologies
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Web Mining and Content Analysis: Large pretrained models with web data

Abstract

With the advent of large language models (LLMs), the vast unstructured text within millions of academic papers is increasingly accessible for materials discovery, although significant challenges remain. While LLMs offer promising few- and zero-shot learning capabilities, particularly valuable in the materials domain where expert annotations are scarce, general-purpose LLMs often fail to address key materials-specific queries without further adaptation. To bridge this gap, fine-tuning LLMs on human-labeled data is essential for effective structured knowledge extraction. In this study, we introduce a novel annotation schema designed to extract generic process-structure-properties relationships from scientific literature. We demonstrate the utility of this approach using a dataset of 128 abstracts, with annotations drawn from two distinct domains: high-temperature materials (Domain I) and uncertainty quantification in simulating materials microstructure (Domain II). Initially, we developed a conditional random field (CRF) model based on MatBERT, a domain-specific BERT variant, and evaluated its performance on Domain I. Subsequently, we compared this model with a fine-tuned LLM (GPT-4o from OpenAI) under identical conditions. Our results indicate that fine-tuning LLMs can significantly improve entity extraction performance over the BERT-CRF baseline on Domain I. However, when additional examples from Domain II were incorporated, the performance of the BERT-CRF model became comparable to that of the GPT-4o model. These findings underscore the potential of our schema for structured knowledge extraction and highlight the complementary strengths of both modeling approaches.

Problem

• Extract process-structure-properties relationships from materials science texts
• Adapt LLMs for materials-specific queries using fine-tuning and annotation
• Compare BERT-CRF and fine-tuned LLM performance on structured knowledge extraction

Innovation

• Fine-tuning LLMs for materials-specific knowledge extraction
• Novel annotation schema for process-structure-properties relationships
• Combining BERT-CRF and GPT-4o for improved performance