2027 — Reproducible and Transparent Machine Learning and AI in Nutrition Research¶
| When | 10 - 15 May 2027 |
| Where | Joachim Herz House, Atlanta, USA |
| Host coordinator | Daniel Kirk, Sander Jannsen, Jan Voges, Mattea Müller |
| Open days | Days 1–3 |
Held together with trainees from the NIH Precision Nutrition Training Programme for transatlantic exchange and alignment of methodological standards. Focus: reproducible ML workflows, FAIR data practices, and explainable AI.
Registration opens in 2026.
To be notified when registration opens, or to ask about attending, contact the coordinators at [add contact address].
Programme¶
Draft programme
Speakers and titles are being confirmed. This outline is indicative.
Day 1 — Predictive modelling of multi-modal nutrition data¶
- Member talks on predictive modelling of multi-modal nutrition data and on the challenges of ML in postprandial studies
- Expert talk on machine-learning methods and theory
- External talk on methods for metabolomics and multi-omics data
- Member talk on clinical validation, from theory to practice
Day 2 — Open science, reproducibility, and making nutrition data FAIR¶
- Keynote on trustworthy system design: robustness, explainability, fairness
- Expert talks on modern machine-learning approaches for massive biomedical data, and on ML for host–bacteria interactions
- Hands-on workshop — Avoiding common pitfalls: hands-on ML in Python with real nutrition datasets: stability and uncertainty versus sample size, repeating validation procedures, feature-selection stability, and the sensitivity of feature importance to the choice of metric
Day 3 — Network working day¶
Reserved for network members.
Venue & travel¶
Placeholder
Venue address, accommodation, and travel information will be added here.