Fourier Neural Operators (FNO) are a reference technique based on neural networks to estimate the solution of differential equations. We propose two tutorial sessions to learn about FNO. The goal is to enable as many people as possible to become aware of, understand and start experimenting with this technique.

Session 1 on Friday May 23 from 2pm to 4:30pm at LMA:

• Method and Architecture (Christophe Gomez)
• Presentation of a paper on seismic (Régis Cottereau)
• Application with a simple code (Valentin Emiya)
• Praticals/homework (individually or in group) : getting into the code with simple changes, adapting the code to a personal use case

Session 2 on Friday, June 27 from 2pm to 4:30pm at LMA:

• TP feedback (10-15 minutes per person/group)
• Discussions on experience feedbacks and perspective

Even though we intend to give a presentation that requires no prerequisites, it may be useful for participants to familiarize themselves with these concepts before the first date, so as to get the most out of the meetings, for example:

• watch the introduction to PINN and FNO available online at https://www.fidle.cnrs.fr/w3/sequences/16-PINNS.html.
• read the reference article on FNOs (https://arxiv.org/abs/2010.08895), quickly or in depth.