The Artificial Intelligence in Health Sciences Symposium took place on April 16 at the Faculty of Medicine and Health Sciences of the University of Barcelona. The event was co-organized by the Faculties of Mathematics and Informatics, Physics and Medicine and Health Sciences, with the support of the Institute of Neurosciences of the University of Barcelona (UBneuro) and funding from the Vice-Rectorate of Research. It gathered clinicians and researchers to explore how AI is transforming medicine and biomedical sciences, while critically addressing its real-world challenges.
The event opened with Dr. Antoni Trilla, dean of the Faculty, who emphasized that AI is already reshaping diagnosis, prediction, and healthcare systems, but must remain grounded in scientific rigor, clinical relevance, and ethical responsibility.
Dr. Petia Radeva (UBneuro, also part of the organization) introduced key AI concepts—machine learning, deep learning, and generative AI—highlighting their growing role in early detection, personalized medicine, and digital health monitoring. She also addressed major concerns such as bias, interpretability, data privacy, and environmental impact, framing AI as a tool whose value depends on its governance.
The keynote lecture was delivered by Ninon Burgos, CNRS research director at the Paris Brain Institute, who presented her work on generative models for anomaly detection in neuroimaging. Her approach is based on training algorithms using images of healthy brains to learn their underlying distribution, and subsequently identifying anomalies through deviations in reconstructed images. The results demonstrated the potential of these models to provide clinically meaningful indicators and to simulate different stages of disease progression.
The programme also included several short oral presentations. Simone Balocco and Berta Alegre (UB and Hospital Clínic) presented a collaborative project between clinicians and engineers to improve the segmentation of head and neck tumors, highlighting how AI can enhance precision in complex imaging tasks. In neuroscience, Jordi Abante (UBneuro) explored AI approaches to identify genetic modifiers in neurodegenerative diseases and to analyze large-scale neural population data with cellular resolution.
From a language technology perspective, Martin Krallinger (Barcelona Supercomputing Center) discussed the development of multilingual natural language processing systems tailored for clinical environments, focusing on their evaluation and potential to structure and extract information from medical texts.
Innovation and translational research were also key themes. Agustín Gutiérrez-Gálvez (UB) presented AI-driven metabolomics approaches for precision diagnostics, combining biochemical data, advanced instrumentation, and machine learning to identify biomarkers. Jordi Pegueroles (Ephion Health) showcased how gait analysis can be used to derive digital biomarkers, enabling non-invasive, continuous monitoring for preventive healthcare.
Finally, Guillem Anglada (European Research Council Executive Agency) introduced advanced language models designed to process and understand scientific content, with applications in screening and classifying the large volume of proposals submitted to ERC calls.
The symposium concluded with a talk by the UB International Research Projects Office (OPIR), who provided an overview of research programmes and calls. An open discussion followed, giving researchers space to build synergies and collaborations while identifying future research directions. Overall, the event reinforced the University of Barcelona’s role as a hub for interdisciplinary dialogue at the intersection of AI and health, contributing to a shared vision of more predictive and personalized healthcare.
