MeXtract: Light-Weight Metadata Extraction from Scientific Papers

2025-10-08

AI Summary

To address weak cross-domain generalization and poor schema adaptability in scientific paper metadata extraction, this paper proposes MeXtract—a family of lightweight language models (0.5B–3B parameters) fine-tuned from Qwen-2.5, incorporating schema-aware training and a transferable architecture. We introduce MOLE-Bench+, an extended benchmark specifically designed for metadata extraction, featuring new multi-domain and multi-format test subsets. Experiments demonstrate that MeXtract achieves state-of-the-art performance on MOLE, significantly outperforming existing methods, while maintaining strong generalization to unseen schemas and cross-domain settings. All code, data, and models are publicly released.

Technology Category

• Machine Learning: Mixture of Experts (MoE)
• Natural Language Processing: Information Extraction
• Search and Optimization: Metareasoning and Metaheuristics

Application Category

• Web Mining and Content Analysis: Large pretrained models with web data
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Search and Retrieval-Augmented AI: Multilingual and cross-lingual Web search

Abstract

Metadata plays a critical role in indexing, documenting, and analyzing scientific literature, yet extracting it accurately and efficiently remains a challenging task. Traditional approaches often rely on rule-based or task-specific models, which struggle to generalize across domains and schema variations. In this paper, we present MeXtract, a family of lightweight language models designed for metadata extraction from scientific papers. The models, ranging from 0.5B to 3B parameters, are built by fine-tuning Qwen 2.5 counterparts. In their size family, MeXtract achieves state-of-the-art performance on metadata extraction on the MOLE benchmark. To further support evaluation, we extend the MOLE benchmark to incorporate model-specific metadata, providing an out-of-domain challenging subset. Our experiments show that fine-tuning on a given schema not only yields high accuracy but also transfers effectively to unseen schemas, demonstrating the robustness and adaptability of our approach. We release all the code, datasets, and models openly for the research community.

Problem

• Extracting metadata accurately from scientific papers
• Overcoming domain generalization limitations in extraction models
• Handling schema variations in metadata extraction tasks

Innovation

• Lightweight models fine-tuned from Qwen 2.5
• Achieves state-of-the-art on MOLE benchmark
• Generalizes effectively to unseen metadata schemas