Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

GLaMoR: Consistency Checking of OWL Ontologies using Graph Language Models

Justin Mücke, Ansgar Scherp

TL;DR

This work tackles the scalability challenge of ontology consistency checking in OWL by reframing the task as a graph-structured binary classification using Graph Language Models (GLMs). It introduces GLaMoR, a pipeline that modularizes OWL ontologies via OAPT, injects inconsistencies through anti-patterns, translates modules into triples, and trains GLMs with global attention to detect inconsistencies. On NCBO BioPortal ontologies, GLMs achieve up to $95.13\%$ accuracy and demonstrate a substantial runtime advantage over semantic reasoners like HermiT (≈$122$ hours), while outperforming embedding-based and other baselines. The study highlights the robustness and scalability of global GLMs for ontology reasoning and outlines paths for handling longer sequences and real-world inconsistencies in future work.

Abstract

Semantic reasoning aims to infer new knowledge from existing knowledge, with OWL ontologies serving as a standardized framework for organizing information. A key challenge in semantic reasoning is verifying ontology consistency. However, state-of-the-art reasoners are computationally expensive, and their efficiency decreases as ontology sizes grow. While classical machine learning models have been explored for consistency checking, they struggle to capture complex relationships within ontologies. Large language models (LLMs) have shown promising results for simple reasoning tasks but perform poorly on structured reasoning. The recently introduced Graph Language Model (GLM) offers a way to simultaneously process graph-structured data and text. This paper proposes GLaMoR (Graph Language Model for Reasoning), a reasoning pipeline that transforms OWL ontologies into graph-structured data and adapts the GLM architecture for consistency checking. We evaluate GLaMoR on ontologies from the NCBO BioPortal repository, converting them into triples suitable for model input. Our results show that the GLM outperforms all baseline models, achieving $95\%$ accuracy while being 20 times faster than classical reasoners. The Code is accessible under: https://github.com/JustinMuecke/GLaMoR

GLaMoR: Consistency Checking of OWL Ontologies using Graph Language Models

TL;DR

This work tackles the scalability challenge of ontology consistency checking in OWL by reframing the task as a graph-structured binary classification using Graph Language Models (GLMs). It introduces GLaMoR, a pipeline that modularizes OWL ontologies via OAPT, injects inconsistencies through anti-patterns, translates modules into triples, and trains GLMs with global attention to detect inconsistencies. On NCBO BioPortal ontologies, GLMs achieve up to accuracy and demonstrate a substantial runtime advantage over semantic reasoners like HermiT (≈ hours), while outperforming embedding-based and other baselines. The study highlights the robustness and scalability of global GLMs for ontology reasoning and outlines paths for handling longer sequences and real-world inconsistencies in future work.

Abstract

Semantic reasoning aims to infer new knowledge from existing knowledge, with OWL ontologies serving as a standardized framework for organizing information. A key challenge in semantic reasoning is verifying ontology consistency. However, state-of-the-art reasoners are computationally expensive, and their efficiency decreases as ontology sizes grow. While classical machine learning models have been explored for consistency checking, they struggle to capture complex relationships within ontologies. Large language models (LLMs) have shown promising results for simple reasoning tasks but perform poorly on structured reasoning. The recently introduced Graph Language Model (GLM) offers a way to simultaneously process graph-structured data and text. This paper proposes GLaMoR (Graph Language Model for Reasoning), a reasoning pipeline that transforms OWL ontologies into graph-structured data and adapts the GLM architecture for consistency checking. We evaluate GLaMoR on ontologies from the NCBO BioPortal repository, converting them into triples suitable for model input. Our results show that the GLM outperforms all baseline models, achieving accuracy while being 20 times faster than classical reasoners. The Code is accessible under: https://github.com/JustinMuecke/GLaMoR
Paper Structure (30 sections, 2 figures, 7 tables)

This paper contains 30 sections, 2 figures, 7 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. Web Ontology Language
  4. Ontology Reasoning and Modularization
  5. Embedding Models for Ontologies
  6. Graph Foundational Models
  7. Machine Learning Reasoning
  8. GLaMoR Pipeline
  9. Modularization
  10. Modification
  11. Translation
  12. Embedding
  13. Training
  14. Experimental Apparatus
  15. Dataset
  16. ...and 15 more sections

Figures (2)

  • Figure 1: GLaMoR pipeline of acquiring and preprocessing the data and using it to train the graph language model (GLM).
  • Figure 2: Distribution on the number of ontologies that can be separated into a certain number of modules.