Thematic program - Statistical Physics & Machine Learning

Machine Learning methods will likely become new powerful tools to analyse and understand physics. Reciprocally, methods and ideas developed in statistical physics can play a major role in developing the theory of modern machine learning algorithms. This program will be organized around four thematic weeks (workshops) centred on different topics at the intersection of Physics and Machine Learning. Two days of invited lectures will inaugurate each week. A colloquium will present the theme of the workshop to a wide audience. The second part of each week will be dedicated to discussions between experts.

How, When & Where


Steering Commitee

  • Giulio Biroli, Ecole Normale Supérieure, Paris
  • Marylou Gabrié, Ecole Polytechnique, Palaiseau
  • Remi Monasson, Ecole Normale Supérieure, Paris
  • Levent Sagun, FAIR, Paris



  • Workshop 1 : information to come
  • Workshop 2 : 3-7 October 2022
  • Workshop 3 : information to come
  • Workshop 4 : 7-11 November 2022



  • Conference at the Collège de France (Ulm site)
  • Organized discussions at the ENS-PSL Data Science Center (CSD)
  • Symposium at ENS-PSL


Weekly Schedule

  • Monday and Tuesday: Conference
  • Wednesday, Thursday and Friday: Organized discussions
  • Thursday: Symposium



  • Conference for a large specialized audience
  • Organized discussions for invited speakers and selected participants
  • Symposium for a wide audience


Workshop Themes

  1. Energy Savvy Computation for Artificial and Biological Neural Networks (more information to come)
  2. Machine Learning-Assisted Sampling for Scientific Computing – Applications in Physics 
  3. Statistical Physics for Machine Learning Meets Computational Social Science (more information to come)
  4. Machine Learning Glassy Dynamics