Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Combining Machine Learning and Ontology: A Systematic Literature Review

Sarah Ghidalia, Ouassila Labbani Narsis, Aurélie Bertaux, Christophe Nicolle

TL;DR

This systematic literature review analyzes 128 studies on integrating machine learning with ontologies to fuse inductive and deductive AI. It identifies three core hybrid categories—Learning-Enhanced Ontology, Semantic Data Mining, and Learning and Reasoning Systems—and catalogs the ML algorithms, reasoning capabilities, AI themes, and application domains used across these works, highlighting a strong NLP emphasis and a predominance of subsumption-based ontological reasoning. The authors map these categories to established hybrid AI design patterns and discuss alignment with neuro-symbolic taxonomies, while outlining key challenges in ontology expressiveness, explainability, and consistency management for future work. Overall, the review offers a comprehensive state-of-the-art view and actionable directions for building scalable, explainable hybrid AI systems that leverage ontologies for knowledge representation and reasoning.

Abstract

Motivated by the desire to explore the process of combining inductive and deductive reasoning, we conducted a systematic literature review of articles that investigate the integration of machine learning and ontologies. The objective was to identify diverse techniques that incorporate both inductive reasoning (performed by machine learning) and deductive reasoning (performed by ontologies) into artificial intelligence systems. Our review, which included the analysis of 128 studies, allowed us to identify three main categories of hybridization between machine learning and ontologies: learning-enhanced ontologies, semantic data mining, and learning and reasoning systems. We provide a comprehensive examination of all these categories, emphasizing the various machine learning algorithms utilized in the studies. Furthermore, we compared our classification with similar recent work in the field of hybrid AI and neuro-symbolic approaches.

Combining Machine Learning and Ontology: A Systematic Literature Review

TL;DR

This systematic literature review analyzes 128 studies on integrating machine learning with ontologies to fuse inductive and deductive AI. It identifies three core hybrid categories—Learning-Enhanced Ontology, Semantic Data Mining, and Learning and Reasoning Systems—and catalogs the ML algorithms, reasoning capabilities, AI themes, and application domains used across these works, highlighting a strong NLP emphasis and a predominance of subsumption-based ontological reasoning. The authors map these categories to established hybrid AI design patterns and discuss alignment with neuro-symbolic taxonomies, while outlining key challenges in ontology expressiveness, explainability, and consistency management for future work. Overall, the review offers a comprehensive state-of-the-art view and actionable directions for building scalable, explainable hybrid AI systems that leverage ontologies for knowledge representation and reasoning.

Abstract

Motivated by the desire to explore the process of combining inductive and deductive reasoning, we conducted a systematic literature review of articles that investigate the integration of machine learning and ontologies. The objective was to identify diverse techniques that incorporate both inductive reasoning (performed by machine learning) and deductive reasoning (performed by ontologies) into artificial intelligence systems. Our review, which included the analysis of 128 studies, allowed us to identify three main categories of hybridization between machine learning and ontologies: learning-enhanced ontologies, semantic data mining, and learning and reasoning systems. We provide a comprehensive examination of all these categories, emphasizing the various machine learning algorithms utilized in the studies. Furthermore, we compared our classification with similar recent work in the field of hybrid AI and neuro-symbolic approaches.
Paper Structure (52 sections, 15 figures, 3 tables)

This paper contains 52 sections, 15 figures, 3 tables.

Table of Contents

  1. Introduction
  2. Methodology
  3. Planning the review
  4. Definition of the research questions
  5. Scientific databases and keywords selection
  6. Definition of inclusion and exclusion criteria
  7. Inclusion criteria
  8. Exclusion criteria
  9. Definition of quality criteria
  10. Performing the review
  11. Collecting articles
  12. Applying inclusion, exclusion, and quality criteria
  13. Analysis
  14. Overview of included studies combining ontologies and machine learning techniques
  15. Types of combination between ontology and machine learning (RQ1)
  16. ...and 37 more sections

Figures (15)

  • Figure 1: Systematic Literature Review (SLR) methodology
  • Figure 2: Selection of articles
  • Figure 3: Evolution of the number of publications about the combination of ontologies and machine learning.
  • Figure 4: Country where the first author is based
  • Figure 5: Overview of the combination of ontologies and machine learning techniques
  • ...and 10 more figures