PEPR DIADEM (Diamond), PEPR IA (CAUSALI-T-AI, PDE IA), PEPR NumPex

ORGANISERS

Marianne CLAUSEL and Noël JAKSE

Call for proposal : The Inter-PEPR Day, jointly organized by PEPR DIADEM, PEPR IA and PEPR NumPex aims to offer an overview of some promising research directions at the interface between Material Science, AI and High Performance Computing

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

Program

• 8h30-9h00: Welcoming & Coffee

• 9h00-10h00: Marie-Pierre GAIGEOT (IUF & Evry University) Exploiting graph theory in MD simulations for extracting chemical and physical properties of materials

• 10h-10h40: Émilie DEVIJVER (CNRS & University of Grenoble Alpes) An introduction to causality

• 10h40-11h00: Coffee break

• 11h00-11h40: Guillaume STIRNEMANN (CNRS, ENS) Machine learned interatomic potentials for reactivity in the condensed phases: from recent achievements to future developments

• 11h40-12h20: Stéphane CHRÉTIEN Introduction to conformal prediction

• 12h20-14h00: Lunch break & Posters session

• 14h00-15h00: Amaury HABRARD (IUF & University Jean Monnet) Domain adaptation and transfer learning

• 15h00-15h40: Alessandro LEITE (INRIA) Causal Knowledge Discovery through Large Language Models: Challenges and Opportunities

• 15h40-16h00: Coffee break

• 16h00-16h40: Emmanuel FRANCK (UNISTRA) Hybrid ML approaches in Material Science : the contribution of PINNS

• 16h40-17h20: Martin URHIN (University of Grenoble Alpes) When GenAI meets Material Science

• 17h20- : Discussion and Conclusion