Rethinking OWL Expressivity: Semantic Units for FAIR and Cognitively Interoperable Knowledge Graphs Why OWLs don't have to understand everything they say

Abstract

Semantic knowledge graphs are foundational to implementing the FAIR Principles, yet RDF/OWL representations often lack the semantic flexibility and cognitive interoperability required in scientific domains. We present a novel framework for semantic modularization based on semantic units (i.e., modular, semantically coherent subgraphs enhancing expressivity, reusability, and interpretability), combined with four new representational resource types (some-instance, most-instances, every-instance, all-instances) for modelling assertional, contingent, prototypical, and universal statements. The framework enables the integration of knowledge modelled using different logical frameworks (e.g., OWL, First-Order Logic, or none), provided each semantic unit is internally consistent and annotated with its logic base. This allows, for example, querying all OWL 2.0-compliant units for reasoning purposes while preserving the full graph for broader knowledge discovery. Our framework addresses twelve core limitations of OWL/RDF modeling, including negation, cardinality, complex class axioms, conditional and directive statements, and logical arguments, while improving cognitive accessibility for domain experts. We provide schemata and translation patterns to demonstrate semantic interoperability and reasoning potential, establishing a scalable foundation for constructing FAIR-aligned, semantically rich knowledge graphs.

arXiv admin note: text overlap with arXiv:2301.01227

References in corpus (13)

• Knowledge Graphs
• Unifying Large Language Models and Knowledge Graphs: A Roadmap
• From human experts to machines: An LLM supported approach to ontology and knowledge graph construction
• Iterative Zero-Shot LLM Prompting for Knowledge Graph Construction
• FAIR 2.0: Extending the FAIR Guiding Principles to Address Semantic Interoperability
• Extract, Define, Canonicalize: An LLM-based Framework for Knowledge Graph Construction
• The FAIREr Guiding Principles: Organizing data and metadata into semantically meaningful types of FAIR Digital Objects to increase their human explorability and cognitive interoperability
• Right for Right Reasons: Large Language Models for Verifiable Commonsense Knowledge Graph Question Answering
• Harnessing Large Language Models for Knowledge Graph Question Answering via Adaptive Multi-Aspect Retrieval-Augmentation
• Semantic Units: Organizing knowledge graphs into semantically meaningful units of representation
• A Comprehensive Evaluation of Neural SPARQL Query Generation from Natural Language Questions
• Improving Language Models with Intentional Analysis
• The Grammar of FAIR: A Granular Architecture of Semantic Units for FAIR Semantics, Inspired by Biology and Linguistics