TL;DR
This paper introduces crowd 2.0, a web framework designed to bridge the gap between visual and formal ontology development tools. It enables seamless interchange of ontologies represented in various visual formats (UML, ER, ORM 2) and formal languages (OWL 2), facilitating collaboration between domain experts and ontology engineers.
What Problem Does This Paper Address?
The adoption of Semantic Web technologies, which rely on ontologies for shared understanding, is hindered by a lack of interoperability between the diverse tools used for ontology development and support. Specifically, this paper tackles the challenge of enabling different tools to exchange and manipulate ontologies, which often exist in both visual (human-friendly) and serialised (machine-readable) formats. This is crucial because ontology engineering involves various stakeholders with different skill sets, from domain experts who prefer visual representations to ontology engineers who work with formal logic-based languages. The current landscape suffers from weak support for integrating these disparate tools and formalisms, leading to difficulties in collaboration and potential errors in ontology creation.
Methods
The core of crowd 2.0 is a web framework that unifies standard conceptual data modelling languages (UML, ER, ORM 2) for generating OWL 2 ontologies from semantic visualisations. The key innovation lies in the use of an "ontology-driven metamodel" called KF (Knowledge Formalism). KF acts as a bridge, preserving the semantics of visual models and allowing for interchange between different representations.
The workflow involves:
- Visualisation: Users can create visual models using UML, ER, or ORM 2 diagrams.
- Metamodel Conversion: These visual models are translated into instances of the KF metamodel.
- Formalisation: The KF instances are then formalised into Description Logics (DL) for automated reasoning.
- Reasoning: DL reasoners can be used to detect errors or infer new knowledge.
- Update and Export: Reasoning results can be incorporated back into the KF instance, and the model can then be exported into various visual languages or as an OWL 2 ontology.
The framework is implemented as a web application with a client-server architecture, connecting various applications for database management, KF rule implementation, and interaction with DL reasoners.
Datasets
Not discussed in the paper. The paper focuses on a framework and its application to a conceptual domain (COVID-19 vaccines) rather than on training or evaluating against specific datasets.
Key Insights and Ideas
- KF as a Semantic Bridge: The KF metamodel is the central piece, enabling semantic interoperability by acting as a common ground between diverse visual and formal ontology representations. This decouples the user from the underlying formalisms.
- Independence from Visual Languages and Logic Tools:
crowd 2.0aims to provide independence from specific visual languages and logic-based tools, allowing users to choose their preferred tools while still benefiting from automated reasoning and interoperability. - User-Centred Ontology Engineering: The framework is explicitly designed for user-centred ontology development, catering to the needs of both domain experts (visual tools) and ontology engineers (formal reasoning).
- Scenario-Based Demonstration: The paper effectively illustrates its capabilities through three scenarios involving the COVID-19 vaccine domain, showcasing how different conceptual views can be managed, how conceptualisations can be implemented, and how automated reasoning can be used to update models.
Key Results
The paper demonstrates a functional framework that enables interoperability between UML, ER, and ORM 2 for ontology development. It provides a web service for the KF metamodel and integrates a new ER editor. The framework supports reasoning over models regardless of their visual representation, and the latest version is available online. While specific quantitative results are not detailed, the qualitative outcome is a more integrated and collaborative ontology development process.
Limitations and Future Work
The authors acknowledge that a main limitation is the current inability to import arbitrary OWL 2 ontologies for reuse (visualisation or editing). Future work will focus on embedding a more expressive subset of OWL 2 axioms into KF instances to address this. Additionally, they plan to assess the usability of the tool by evaluating usability factors such as the System Usability Scale (SUS). Further development will also consider increasing the expressiveness of the approach to cover more ontology languages and features.
Summary updated