MathTech Training
The FMJH in collaboration with the EDMH, offers a training program to develop
cross-disciplinary skills, specifically designed for EDMH doctoral students.
Why develop cross-disciplinary skills ?
- To promote what you do by learning how to “sell” yourself better, share what you know, and pass it on
- To open up new opportunities and add interpersonal skills
- Develop your open-mindedness, understand the issues at stake, and identify new uses.
- Create connections, avoid silos, and promote multidisciplinarity.
As part of this program, the FMjh and the EDMH are organizing two key events and two workshops
2025-2026 program
- Decembre 16, 2025 : "Pitch" Workshop
- Janvier 28, 2026 : MathTech meetings, "From a PhD in mathematics to the business world: testimonials"
- April 14, 2026 : Training on "How to leverage our doctorate outside the academic world"
- May 18-22, 2026 : MathTech Challenge.
MathTech challenge 2026
- Date: May 18–22
- Location: Orsay Institute of Mathematics
- Corporate Challenge: Michelin
- Two topics will be offered:
-
Topic 1: Computational mechanics simulation for design at Michelin
Objective: Create a learning model or propose a methodological/theoretical approach to rapidly predict physical fields (deformation, pressure, displacement, etc.) on new geometries, to accelerate tire rolling simulations.
Why: Test dozens of possible geometries in seconds instead of weeks of simulation—more experimentation, more innovation.
Code/data provided: A turnkey code for rolling simulations and all the materials needed to create your own simulation more quickly.
Method ideas: Use operator networks (FNO/DeepONet), graphs/meshes (GNN), or implicit representations (SDF/NNs), or a mix of all three to leverage their strengths.
-
Topic 2: Explainable Industrial Process Data Generator
Objective: To create a realistic data generator based on measurement series and a method for explaining the implicit relationships between parameters. The generator must be able to use both tabular data and time series.
Why: Data augmentation to perform advanced analyses or to train and test AI models. Explanation of the implicit relationships between parameter values.
Data provided: A set of measurements from a large-scale industrial machine (>15). A physics-inspired toy model to independently test and validate methodologies.
The winners of the challenge will each receive €400!
You can register on your ADUM account