SMART Summer School 2025

Last week, our SMART doctoral researchers & postdocs traveled to Paris to participate in a three-day summer school in collaboration with Gustave Roussy Institute & Sorbonne Université. The program focused on an interdisciplinary academic exchange and brought together researchers with backgrounds in biomedicine, statistics, mathematics and computer science.

Day 1 – June 30, 2025

After arrival in Paris, we headed straight to the Gustave Roussy Institute for an afternoon program titled: Multidisciplinary Perspectives on AI and Data Integration in Medicine. The session featured the following presentations from researchers from Gustave Roussy and SmallData:

• Meta-learning for regularizing deep networks under small data
  regimes (Lennart Purucker)
• Longitudinal ctDNA-based model for survival analysis in ER+/HER2-advanced breast cancer (Aaron Mamann)
• Reciprocal transfer of knowledge from population-based genetic screens to whole-body, organ-resolved models of human metabolism (Daniel Fässler)
• Exploring markers of survival and response under immunotherapy in advanced melanoma (Ines Malleval)
• Similarity of individual latent dynamics in longitudinal cohort data (Clemens Schächter)
• AI for Visual Computing and Medical Images (Pierre Marza)
• An effective similarity integration multi-modal graph neural network method to facilitate disease gene prioritization (Adrian Fritz)
• Mathematical modelling in Blood Cancers (Gurvan Hermange)

The diversity of topics and perspectives sparked rich discussions. Networking between groups laid the groundwork for potential inter-institutional collaborations. We wrapped up the day with an authentic french dinner and informal networking in the city.

Day 2 – July 1, 2025

Tuesday was hosted at the campus of Sorbonne Université and kicked off with lab presentations from institutions including LIP6, ISIR, SCAI, LPSM, and SmallData. To energise the room, we played Human Bingo, which encouraged participants to network and connect beyond their fields and own institutes. This was followed by Topic Table Discussions, where moderators facilitated focused conversations on the following themes:

• Benchmarking and AutoML for Scientific Applications (mod: Asif Hasan)
• Synthetic Data & Simulation Studies in Science (mod: Clemens Schächter)
• Principles of Foundation Models (mod: Niklas Brunn)
• Large Language Models (mod: Ryan Köksal)
• Multi-Modal and Heterogeneous Data Integration (mod: Jonatan Menger)
• Cross-Disciplinary Collaboration in Computational Life Sciences (mod: Laia Canal Guitart)
• Personalized & Precision Medicine: Modeling Clinical Data (mod: Daniel Fässler, Max Behrens)
• Uncertainty in Deep Learning (mod: Yaser Kord)

Day 3 – July 2, 2025

On the last day, we wrapped up the program with a Paris Science Quest , a city-based challenge combining sightseeing, cultural and science related trivia. Starting at the Louvre and ending at Notre Dame , participants walked, solved riddles, and rediscovered landmarks through the lens of data. After lunch and a group photo at the Notre Dame, the group made their way back to Freiburg – concluding the Summer School with valuable insights and strengthened networks and team work.

Thanks to all our collaborators, hosts, researchers, and of course, the Deutsche Forschungsgemeinschaft for making this event possible.

Marc Schumacher

Institute of Medical Biometry and Statistics,
Faculty of Medicine and Medical Center –
University of Freiburg