Description
AISSAI/GAP2026 is the second edition of the Grenoble AI for Physical Sciences (GAP) workshop, following the great success of its inaugural edition in 2024. GAP brings together international experts working at the intersection of artificial intelligence and the physical sciences, with a particular focus on inverse problems, simulation-based inference, data-driven discovery, and foundational models for science. The 2026 edition will continue to highlight cutting-edge advances in theory, algorithms, high-performance computing, and applications across domains such as climate and geosciences, astrophysics, neuroscience, and engineering systems. The format emphasizes pedagogical and forward-looking invited talks that introduce fundamental concepts and survey current challenges and research frontiers. We expect around 200 participants this year. In addition to invited talks, the programme will feature a poster session to showcase ongoing research, encourage exchanges among participants, and stimulate new collaborations across disciplines.
Attendance and registration
The key dates for attending the workshop on-site are:
- 17-April 2026 : deadline for poster submissions
- 01-May 2026 : notification of acceptance of poster submissions
- 15-May 2026 : deadline for all registrations
The event will be streamed at https://videos.univ-grenoble-alpes.fr/live/events/.
Keynote speakers
We will have the pleasure of having keynote talks of the following confirmed speakers:
- Linus Bleinstein, EPFL
(foundational models for life sciences) - Patrick Galinari, Sorbonne Université
(foundational models for scientific applications) - Samuel Hurault, CNRS, Université Gustave Eiffel
(diffusion models, flow matching, and inverse problems) - Ching-Yao Lai, Stanford University
(physics-informed deep learning for inverse problems) - Fanny Lehmann, ETH Zurich
(foundational models for geophysical sciences) - Jakob Macke, University of Tuebingen
(simulation-based inference) - Julien Mairal, Inria
(inverse problems for imaging) - Laurence Perrault-Levasseur, Université de Montréal
(bayesian methods for astronomy and cosmology) - Nelly Pustelnik, CNRS, ENS Lyon
(model-based neural networks) - Gabrielle Steidl, TU Berlin
(generative modelling)
All presentations will provide a comprehensive overview of the field, equipping participants with the knowledge they need to delve into the specific applications and potential of machine learning in their respective domains.
Poster session
Program
The program is currently being polished and will soon be available.
However, what can be said already is that:
- The talks on Wednesday 17-June will be from 9h to 12h30 and 14h to 17h30.
-- The workshop will propose a standing buffet for the lunch break, during which the contributed posters will be displayed.
- The talks on Thursday 18-June will be from 9h to 12h30 and 14h to 17h30.
-- The workshop will propose a standing buffet for the lunch break, during which the contributed posters will be displayed.
- The talks on Friday 19-June will be from 9h to 12h30.
Location
Maison de la Création et de l'Innovation (MaCi) @ Université Grenoble Alpes
339 avenue Centrale
Domaine Universitaire
38400 Saint-Martin-d'Hères
Organizers
- Scientific Program and local organization:
- Jérôme Bobin, CEA Saclay
- Julien Le Sommer, CNRS, IGE
- Tobias Liaudat, CEA Saclay
- Thomas Moreau, Inria Saclay
- Bruno Raffin, Inria, LIG
- Pedro L. C. Rodrigues, Inria, LJK
- Administrative support:
- Gaelle Magand (LJK)
Sponsors
