This poster will present the state of KBR’s- the Royal Library of Belgium, multi service data-driven research infrastructure. This is centered on the concept of "Collections as Data," aiming to provide researchers and users with high-quality, machine-readable access to Belgium’s cultural and scientific heritage.

This is supported through three projects that span 2020- 2030, funded by BELSPO. The first step of this infrastructure was the DATA-KBR-BE project under the BRAIN-be 2.0 program. This initiative addressed the technical and legal barriers to data-level access, moving beyond simple online catalogs to provide access to pre-curated digitized collections. This established a sustainable data extraction workflow and the launching the Open Data Platform.

In the second step the data provision services are being further developed through the BELSPO ESFRI project of Virtual Lab, with contributions to DARIAH. In Virtual Lab accredited users of eduGAIN will be able to view available data and make selections for compiling their own datasets. This will be compressed and sent through a data service that further supports BELNETs file transfer services. In a third step the research data infrastructure will be enhanced by the BELSPO ESFRI CultuR-AI project, a contribution to CLARIN. In CultuR-AI KBR will work to evaluate the use of GenAI tools for enriching KBR’s collections. These two infrastructure projects are coordinated by the Digital Research Lab, a 10 year collaboration between KBR and Ghent University, to facilitate text and data mining on KBR’s digitized and born digital collections.

The combination of these projects will aid to facilitate interdisciplinary research, allowing scholars to treat digitized collections and documents as vast datasets for text and data mining (TDM). This also makes clear contributions to the European Research Infrastructures, specifically DARIAH (Digital Research Infrastructure for the Arts and Humanities) and CLARIN (Language Resources and Technology). Through these frameworks, and the FWO CLARIAH-VL initiative, KBR ensures that its data services are interoperable, standardized, and integrated into the broader FEDOSC and European Open Science Cloud (EOSC). This collaboration provides users with accredited systems for secure and persistent access to sensitive or rights-restricted data, fostering a trustworthy environment for large-scale analysis. In addition this works to both streamline internal operations and enhance user experience:

• Data on Demand & Automated Services: This service simplifies the process of compiling, packaging, and sending specific sub-sets of collections to users, reducing the manual burden on library staff while providing tailored datasets for specialized queries.

• Collection Enrichment: AI and machine learning models are utilized for semantic enrichment. This includes automated metadata extraction, Named Entity Recognition (NER), etc.

• User-Level Access: For the end-user, AI-driven interfaces facilitate more intuitive discovery, allowing for similarity searches, across (centuries) of material.

This multi-service infrastructure not only automates internal workflows for data delivery but also empowers the global research community to unlock new insights from Belgium’s rich heritage through sophisticated, data-level exploration.