From Subsumption to Satisfiability: LLM-Assisted Active Learning for OWL Ontologies

2026-04-17

AI Summary

This work addresses the inefficiency and limited scalability of traditional ontology-based active learning, which relies heavily on expert judgments for membership queries. It proposes a novel approach that integrates large language models (LLMs) as a third-party component in the active learning pipeline. Specifically, subsumption relations are reduced to satisfiability problems, and controlled natural language prompts are employed to generate realistic counterexamples for negative concepts, thereby tolerating only Type II errors while preserving ontological consistency. The method’s robustness and practicality are empirically validated across 13 prominent LLMs and multiple standard ontologies, demonstrating stable recall performance and significantly enhancing the scalability and automation of ontology learning.

Technology Category

• Machine Learning: Active Learning
• Knowledge Representation and Reasoning: Ontologies
• Constraint Satisfaction and Optimization: Constraint Learning and Acquisition

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
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs

Abstract

In active learning, membership queries (MQs) allow a learner to pose questions to a teacher, such as ''Is every apple a fruit?'', to which the teacher responds correctly with yes or no. These MQs can be viewed as subsumption tests with respect to the target ontology. Inspired by the standard reduction of subsumption to satisfiability in description logics, we reformulate each candidate axiom into its corresponding counter-concept and verbalise it in controlled natural language before presenting it to Large Language Models (LLMs). We introduce LLMs as a third component that provides real-world examples approximating an instance of the counter-concept. This design property ensures that only Type II errors may occur in ontology modelling; in the worst case, these errors merely delay the construction process without introducing inconsistencies. Experimental results on 13 commercial LLMs show that recall, corresponding to Type II errors in our framework, remains stable across several well-established ontologies.

Problem

• active learning
• OWL ontologies
• subsumption
• satisfiability
• Large Language Models

Innovation

• active learning
• OWL ontologies
• subsumption to satisfiability
• large language models
• controlled natural language