Towards Cross-Domain Linking of Data: A Semantic Mapping of Cultural Heritage Ontologies

The Linked Open Vocabularies (LOV) registry, designed with the Linked Data principles at core, provides an environment suitable for research which targets domain-specific, but also potentially reusable, information representation. The main purpose of this study is to follow the recommendations pertaining to the utilisation of LOV as a basis for experimentation in order to examine how information within the Cultural Heritage (CH) domain can be improved in terms of reusability and interoperability. The present lack of cross-domain knowledge transfer forms the motivation behind this study, with the aim of facilitating the transition from conventional, domain-specific knowledge representation to reusable and semantically interoperable information. The methodology of this study involves the manual semantic mapping of elements from 12 vocabularies in the LOV registry, reinforced by a small-scale experiment using contemporary large language models (LLMs), particularly GPT, for a preliminary assessment of the mapping process. The findings revealed several key aspects to consider regarding the alignment of semantically adjacent vocabulary elements in the CH domain and beyond, emphasising the potential unveiled by linking domain-focused schemata to standardised, established ones while preserving the conceptual hierarchies inherent to each individual knowledge domain. The contribution of this research pertains to the vision of linking data across different domains by initiating the alignment among representation schemata in CH, with the ultimate aim to expand beyond the boundaries of the in-word knowledge domain, while employing combinatory methodological approaches of technological means and human expertise to facilitate this process.