Advancing FAIR Digital Objects for Machine-Actionable Research: Integrating Semantic Enrichment in Research Object Ecosystems

Achieving higher levels of FAIR-ness for research artefacts demands not only structured packaging but also semantic enrichment that links textual resources to knowledge bases. ROHub, a reference platform implementing the Research Object paradigm, enables scientists to package and share research outputs as structured Research Object Crates (RO-Crates) - combining data, methods, software, and associated metadata into a unified, machine-processable entity. While RO-Crates inherently improve metadata richness and FAIR compliance by aggregating diverse resources with persistent identifiers and schema-based annotations, many research outputs still contain unlinked textual artefacts (e.g., reports, questionnaires, narratives) whose contextual semantics remain underutilized. Manual semantic annotation to link these textual elements to external knowledge bases - such as domain ontologies or vocabularies - is time-consuming and error-prone, yet crucial for enhancing findability, semantic interoperability, and machine-actionability. To address this gap, we extend ROHub with an automated semantic annotation service that identifies entities within text resources and links them to relevant knowledge bases, producing enriched metadata that feeds back into the RO-Crate structure. This service integrates entity linking techniques to reduce manual curation overhead and systematically increase the FAIRness and discoverability of research objects - making them more accountable to machine discovery, integration, and automated workflows. The result is a FAIR research object ecosystem where textual content, semantic context, and structured metadata co-exist in a machine-processable form, enhancing both human and computational reuse.