Rudy Morel

Rudy Morel

Research Fellow

Flatiron Institute

I’m a Research Fellow in the Center for Computational Mathematics of the Flatiron Institute in New York, and member of the Polymathic AI collaboration.

I work in AI&Science, focusing on building AI models — such as transformers — for predicting spatio-temporal systems in Physics. I’m also interested in the generation and prediction of stochastic time processes, commonly encountered in Finance.

In 2023, I completed my PhD from École Normale Supérieure on compact models of multi-scale processes. I’m deeply grateful to my PhD advisor, Prof. Stéphane Mallat for this incredible journey, as well as Prof. Jean-Philippe Bouchaud for his exceptional guidance.

From 2019 to 2023, I had the privilege of being a member of the EconophysiX chair hosted by CFM and Polytechnique. In 2022, I was delighted to be a visiting PhD student at Rice University in Prof. Maarten de Hoop’s team.

Email: rmorel [at] flatironinstitute [dot] org

Interests

  • pretrained models
  • multi-scale data
  • transformer models
  • path shadowing

Education

  • PhD in Mathematics, 2020-2023

    École Normale Supérieure

  • MS in Probability and Finance, 2018-2019

    Sorbonne Université

  • BSc in Mathematics, 2015-2018

    École Normale Supérieure of Rennes

News

May 2025. Our paper DISCO got accepted at ICML2025!

May 2025. I presented DISCO at the Foundation Model for Science conference.

Apr 2025. I was in Palo Alto for the Foundation Models of the Physical Universe symposium organized by Foresite Labs.

December 2024. I gave a talk in the M2LInES collaboration. Thank you Chris Pedersen and Laure Zanna for the invitation!