Workshop AI 4 Maths
on Using AI to Advance Mathematics
Description
The contribution of mathematics to AI is undeniable. But by reversing the paradigm, could AI itself be used to advance the field of mathematics? What can AI bring to mathematical research, and how?
Recognizing the importance of these questions for their shared discipline, three learned societies — the French Statistical Society (SFdS), the French Society for Applied and Industrial Mathematics (SMAI), and the French Mathematical Society (SMF) — have joined forces to organize a one-day workshop. Its aim is to familiarize mathematicians with the new capabilities of AI agents (particularly in formal language) and to foster discussion and exchange on these topics.
Several perspectives and works will be presented, covering both the applications of mathematics and the use of AI in fundamental research, by academics as well as researchers from the tech industry. The day will conclude with a roundtable discussion broadening the debate to issues of governance, frameworks, and ethics raised by the introduction of these new technologies.
The workshop is open to mathematicians working in the public or private sector, whether established professionals or in training. It will take place in person at the Institut Henri Poincaré in Paris. Participation is free of charge, but registration is required.
Registration deadline: November 10, 2025
Workshop Program
Morning
9:00–9:15 — Welcome
9:15–10:00 — Opening Perspective — Amaury Hayat (Professor, ENPC – IPP)
10:00–11:00 — Why Explain Mathematics to Computers: An Introduction to LEAN — Patrick Massot (Professor, Laboratoire de Mathématiques d'Orsay, Université Paris-Saclay)
11:00–11:30 — Coffee Break
11:30–12:30 — Applying Language Models to Mathematical Problem Solving: From Olympiads to Research — Yann Fleureau (Engineer and Entrepreneur, Co-founder of the Numina Project)
Lunch break (free)
Afternoon
14:00–15:00 — Finding Interesting Mathematical Constructions with Machine Learning Tools — Adam Zsolt Wagner (Google DeepMind)
15:00–15:30 — Coffee Break
15:30–16:15 — From Neural Networks to Solvers: Scalable Physics-Informed Learning for PDEs — Victorita Dolean-Maini (Professor, TU Eindhoven)
16:15–17:00 — François Charton (Research Engineer, Axiom)
17:00–17:30 — Roundtable Discussion: What Are the Consequences (Framework, Governance, Ethics, etc.) of Introducing These New Technologies?
Moderated by Gérard Biau (Professor, Sorbonne University, Director of SCAI)
Participants: