Terms as defined in the Part B proposal and the task register — not paraphrased from memory. Where a term names an external initiative rather than something internal to AI4Access, that's noted.
- Euro-BioImaging ERIC
- The legal entity coordinating AI4Access via its central coordination unit (the "Statutory Seat") in Turku, Finland.
- Bio-Hub (EMBL)
- Euro-BioImaging section coordinating biological imaging activities: user access, training, Node engagement, and general Image Data Services. Home of the knowledge-engineering function this site is written from.
- Med-Hub (CNR)
- Euro-BioImaging section coordinating medical imaging activities: services, training, Node engagement, and support to image data management.
- Node
- A national Euro-BioImaging consortium coordinating several facilities within a country. 39 Nodes in total.
- Facility
- A local provider — an instrument, platform or lab — within a Node. 295 facilities in total.
- Research Navigator ("Navigator")
- The AI assistant developed in AI4Access that routes a stated research problem to suitable technologies, services and workflows.
- INFRA-SERV
- Thematic access-integration Horizon projects where Euro-BioImaging is a partner — named examples: ISIDORe, canSERV, AgroSERV. T2.5 builds the crosswalks between their catalogues and the AI4Access model.
- EOSC LSR
- The European Open Science Cloud's Life Science Research Node — a target platform for exposing AI4Access services under T3.4.
- EUCAIM
- The EUropean Federation for CAncer IMages: a federated European infrastructure of FAIR, anonymised cancer imaging and clinical data, built to connect distributed data holders across Europe under Europe's Beating Cancer Plan.
- UNCAN.eu
- The EU's federated cancer research data hub ("UNderstand CANcer"), connecting National Cancer Data Nodes in alignment with the European Health Data Space.
- EDAM / EDAM-Bioimaging
- A community ontology of bioimage analysis operations, topics, data types and formats. The vocabulary T3.2 maintains and extends for describing analysis services.
- REMBI
- Recommended Metadata for Biological Images — community guidelines for FAIR metadata on microscopy and bioimage data, referenced directly in T2.1's methodology.
- PID (ORCID / ROR / DOI / PIDINST)
- Persistent identifiers forming the knowledge base's linking layer: ORCID for people, ROR for organisations, DOI for outputs and KB releases, PIDINST for individual instruments.
- Data Use Ontology
- A controlled vocabulary for expressing access conditions on a service or dataset, used in T2.1 to model who may use what and under which terms.
- RAG / CAG
- Retrieval-Augmented Generation / Cache-Augmented Generation — the reasoning pattern behind the AI Agent backend (WP4), grounding LLM responses in the ontology-aligned knowledge base rather than model memory alone.
- MCP (Model Context Protocol)
- The adapter layer connecting the AI Agent to external tools and platforms (EOSC, EUCAIM, UNCAN, BioImage Archive) so it can discover, call and chain them during a conversation.
- Hypha / BioEngine
- The agent-orchestration framework from the AI4Life project, named as direct prior experience for the WP4 AI Agent backend's tool-integration design.
- LS-AAI / Life Science Login
- The authentication system the Navigator integrates with for personalised, multi-turn user sessions on the Access Portal.
- SAB
- Euro-BioImaging's Scientific Advisory Board — its evaluation and reporting cycle is one of the things the T2.4 update protocol has to align with, not duplicate.
- KER
- Key Exploitable Result — WP7's term for a project output with a defined path to adoption or exploitation beyond the grant.
- DEC
- Dissemination, Exploitation & Communication — the umbrella strategy WP7 runs, split into four audience tracks (scientific/RI, technical, SME/industry, policy).
- ESFRI
- The European Strategy Forum on Research Infrastructures. Its Roadmap is cited as the frame AI4Access's RI-consolidation outcome answers to.