Schema as Parameterized Tools for Universal Information Extraction

2025-06-02

AI Summary

Universal Information Extraction (UIE) faces dual challenges in in-context learning: inflexible pre-defined schema selection and insufficient dynamic schema generation capability. To address these, we propose Schema-as-Tool, the first framework that formalizes schemas as parameterizable, LLM-callable tools—enabling unified support for closed-, open-, and on-demand IE tasks. Our method integrates tool invocation, schema retrieval, slot-filling-based extraction, and dynamic schema synthesis, synergistically combining in-context learning with lightweight fine-tuning. Evaluated across four IE task categories, the framework achieves adaptive task switching, high schema retrieval accuracy, and extraction performance competitive with LoRA and state-of-the-art UIE systems—while reducing trainable parameters by a significant margin.

Technology Category

• Natural Language Processing: Information Extraction
• Machine Learning: Large Multimodal Models (LMMs)
• Data Mining & Knowledge Management: Conversational Systems for Recommendation & Retrieval

Application Category

• Search and Retrieval-Augmented AI: Search Tool Learning with LLM: Teaching LLMs to invoke search and make use of retrieved information
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Universal information extraction (UIE) primarily employs an extractive generation approach with large language models (LLMs), typically outputting structured information based on predefined schemas such as JSON or tables. UIE suffers from a lack of adaptability when selecting between predefined schemas and on-the-fly schema generation within the in-context learning paradigm, especially when there are numerous schemas to choose from. In this paper, we propose a unified adaptive text-to-structure generation framework, called Schema as Parameterized Tools (SPT), which reimagines the tool-calling capability of LLMs by treating predefined schemas as parameterized tools for tool selection and parameter filling. Specifically, our SPT method can be applied to unify closed, open, and on-demand IE tasks by adopting Schema Retrieval by fetching the relevant schemas from a predefined pool, Schema Filling by extracting information and filling slots as with tool parameters, or Schema Generation by synthesizing new schemas with uncovered cases. Experiments show that the SPT method can handle four distinct IE tasks adaptively, delivering robust schema retrieval and selection performance. SPT also achieves comparable extraction performance to LoRA baselines and current leading UIE systems with significantly fewer trainable parameters.

Problem

• Adapting predefined schemas in universal information extraction
• Unifying closed, open, and on-demand IE tasks
• Improving schema retrieval and selection performance

Innovation

• Schema as Parameterized Tools (SPT) framework
• Unifies closed, open, on-demand IE tasks
• Robust schema retrieval and selection performance