Developing a model for a Data Lab for Science linking with EOSC and RAISE (RAISE pilot) (CSA) -

Expected Outcome:

Project are expected to contribute to the following outcomes:

• a sustainable operational model and reference blueprint for a Data Lab for Science linking with the European Open Science Cloud (EOSC) and the Resource for AI Science in Europe (RAISE), enabling researchers to accelerate the uptake of AI in Science;
• easier access for researchers to large amounts of AI-ready research data, and to associated high-quality data services to seamlessly develop and train AI models for science; and
• better alignment, interoperability and scaling across countries and disciplines of AI workflows and related services such as data pooling, anonymisation, curation, or synthetic data generation.

Scope:

Data Labs are introduced in the Data Union Strategy as a connector between AI factories and the Common European Data Spaces, providing specialised services to enable turning data into useable resources for AI training and the development of AI applications such as data pooling, curation, pseudonymisation, and synthetic data generation.

A significant number of tools and services within the scope of the Data Lab concept have been developed by research infrastructures and scientific service providers, notably organisations that operate or contribute resources to EOSC, the Common European Data Space for research and innovation.

The development of Data Labs addressing the needs of scientists is a key element in the AI in Science strategy, as a collaboration between RAISE and EOSC. This action aims to anchor the implementation of the Data Lab model with the development of EOSC and RAISE. The Data Lab for science will be a key deliverable and a key element of the data work strand of RAISE.

Proposals should include all of the following activities:

• Identify, consolidate and orchestrate existing data, services and other relevant results of Horizon Europe EOSC-related projects, as well as other community-endorsed promising solutions in view of setting up and implementing a Data Lab for Science, including a mapping of such results and solutions to facilitate their navigation, access and use. The results and solutions may encompass relevant data services, but also related policies, guidance and training approaches.
• Identify and collect necessary elements for seamlessly integrating a Data Lab for Science into the existing European data and high-performance computing infrastructures enabling cohesive workflows.
• Identify technical, operational, legal, organisational or any other type of bottlenecks or gaps preventing the seamless sharing, reuse or pooling of data for AI development and applications in science.
• Develop a blueprint for a Data Lab for Science addressing such bottlenecks and support its implementation through the demonstration of use cases, in coordination with the EOSC Federation and relevant developments in the EuroHPC Federation.
• Develop governance and financing models that ensure a sustainable pathway for a Data Lab for Science beyond the project’s duration.
• Identify researchers’ needs and gaps on data services in different scientific disciplines that are not addressed by existing solutions and propose recommendations for future policy and funding actions.
• Connect and encourage collaboration among projects offering data and data-related services for AI in Science.

The action should build on relevant results of Horizon Europe projects that contribute to the development of EOSC and RAISE, as well as EU-funded projects developing AI solutions for science such as foundation models in science, data commons, or AI research assistants.

The action should closely coordinate with relevant Horizon Europe projects such as those funded under HORIZON-INFRA-2025-01-EOSC-03, HORIZON-CL4-INDUSTRY-2025-01-DIGITAL-62 and HORIZON-CL4-INDUSTRY-2025-01-DIGITAL-61, as well as with relevant projects under EuroHPC and the AI Factories.

News flashes