Several areas of collaboration with the AI4EOSC project were explored during AI4Life’s third General Assembly, held on October 21-22, 2024, in Milan, Italy.
Throughout the event, AI4EOSC contributed actively to discussions on improving interoperability and reinforcing FAIR data principles. A key area of collaboration is the integration of AI4Life Zoo models within the AI4EOSC platform, as the technologies behind both projects are highly compatible, allowing for seamless integration with minimal effort.
AI4EOSC’s presentation emphasised the project’s ongoing developments and the current and future collaborations, particularly for integrating AI4Life’s contributions into larger contexts like the European Open Science Cloud (EOSC). This collaboration promises to open new avenues for sustainability and reusability in the field of AI and open science.
A central theme of the assembly was the importance of FAIR data practices. Key challenges discussed included annotation integration, data visualization, and streamlined debugging. AI4EOSC proposed leveraging established resources, such as the FAIR4ML group at the Research Data Alliance (RDA), and stressed the importance of clear data identifiers and provenance standards. Further discussions also focused on enhancing reproducibility using frameworks like Prov-O and ontologies, reinforcing AI4EOSC’s commitment to advancing reliable, open science methodologies.
The event presented significant advancements in model functionality and the development of the BioImage Model Zoo (BMZ), AI4Life’s catalogue dedicated to creating, uploading, and consuming models. Demonstrations highlighted tools such as BiaPy, DeepImageJ, and Ilastik, which showcased the application of these models in inference tasks.
The assembly also focused on project updates, community engagement, and fostering collaboration across European research infrastructures.
