The Inter-PEPR Day, organized jointly by PEPR DIADEM, PEPR IA and PEPR NumPex, aims to provide an overview of some promising research directions at the interface between materials science, AI and high-performance computing.

The day is devoted to several lectures given by renowned researchers and specialists in the field. Young researchers are also invited to present their work through posters or short talks, and must submit an abstract to inter-pepr-days@services.cnrs.fr by October 15, 2024.

• 8:30-9am: Welcome coffee
• 9-10am: Marie-Pierre GAIGEOT (Université d'Evry) Graphs and numerical simulation
• 10-10:40am: Émilie DEVIJVER (CNRS & Université Grenoble Alpes) Introduction to causality
• 10:40-11am: Coffee break
• 11-11:40am: Guillaume STIRNEMANN (CNRS, ENS) Machine-learned interatomic potentials for condensed-phase reactivity: from recent achievements to future developments
• 11:40-12:20pm: Stéphane CHRÉTIEN Introduction to conformal prediction
• 12:20-2pm: Lunch break and poster session
• 2-3pm: Amaury HABRARD (IUF & Université Jean Monnet) Domain adaptation and transfer learning
• 3-3:40pm: Free time (TBA)
• 3:40-4pm: Coffee break
• 4-4:40pm: Emmanuel FRANCK (UNISTRA) Hybrid ML approaches in materials science: the contribution of PINNS
• 4:40-5:20pm: Martin URHIN (Université Grenoble Alpes) When GenAI meets materials science