Challenges and Opportunities of LLM-Augmented Semantic Model Creation for Dataspaces

The objective of dataspaces is to facilitate seamless and reliable data exchange between different organizations. In Europe, their prominence has grown with the implementation of the European Data Governance Act. This legislation emphasizes the importance of trust, accessibility, and shared dataspaces, which necessitate semantic interoperability grounded in the FAIR principles. While semantic descriptions in the form of semantic models and ontologies are indispensable to dataspaces, their full potential remains unexploited. Meaningful metadata, including contextual information, enhances data usability, but the manual creation of semantic models can be challenging. Large Language Models (LLMs) offer a new way to utilize data in dataspaces. Their advanced natural language processing capabilities enable context-aware data processing and semantic understanding. This paper presents initial experiments on customizing and optimizing LLMs for semantic labeling and modeling tasks. The contributions of this work include research questions for future investigations, early experiments demonstrating the applicability of LLM for semantic labeling, and proposed directions to address discovered challenges.