PhD student at CERMICS (Ecole Nationale des Ponts et Chaussées) under the supervision of Gabriel Stoltz.
My focus is on the understanding of statistical physics and random systems from a mathematical perspective, both concerning theoretical and practical aspects. To be more specific, my PhD is concerned with designing efficient algorithms for sampling rare events probabilities in random systems, such as Feynman-Kac dynamics or adaptive algorithm. I have also been involved in a Machine Learning project for statistical physics in collaboration with the French Nuclear Agency and the Los Alamos National Laboratory, and I am still interested in the links between statistical physics, high dimensional sampling and Machine Learning.
From a broader perspective, I am convinced that developing the interactions between science, industry and politics is a crucial tool in order to face upcoming challenges.
- Partial differential equations, functional inequalities
- Markov chains, Feynman-Kac models and their stability
- Numerical analysis of SDE's
- Stochastic algorithms (MCMC, stochastic approximation, sequential Monte Carlo...)
- Interaction between Machine Learning and statistical physics
About the ergodicity of Feynman-Kac semigroups, with M. Rousset and G. Stoltz
Simulating log-gases from random matrix theory using Hybrid Monte Carlo algorithms, with D. Chafaï.
Ferré, G. & Touchette, H. (2018). Adaptive sampling of large deviations. Preprint
Ferré, G, & Stoltz, G. (2017). Error estimates on ergodic properties of Feynman-Kac semigroups. Preprint
Machine Learning and model reduction
Ferré, G., Haut, T., & Barros, K. (2017). Learning molecular energies using localized graph kernels. The Journal of Chemical Physics, 146(11), 114107. Preprint
Zentner, I., Ferré, G., Poirion, F., & Benoit, M. (2016). A biorthogonal decomposition for the identification and simulation of non-stationary and non-Gaussian random fields. Journal of Computational Physics, 314, 1-13. Paper
Ferré, G., Maillet, J. B., & Stoltz, G. (2015). Permutation-invariant distance between atomic configurations. The Journal of Chemical Physics, 143(10), 104114. Preprint
Talks in conferences
IPAM, UCLA, Los Angeles, Understanding Many-Particle Systems with Machine Learning, 2016, Young Researchers’ Talk. Learning potential energy landscapes with localized graph kernels. Slides
ICTS, Bengalore, Large deviations theory in statistical physics, recent advances and future challenges, 2017, Error estimates for Feynman-Kac semi-groups. Video
IHP, Paris, Young Probabilists' Day (Les probabilités de demain), 2018, Long time stability of Feynman-Kac models. Slides
Talks in seminars
IHP, Paris, Stochastic Dynamics Out of Equilibrium, 2017, Young Researchers’ Seminar. Error estimates for Feynman-Kac semi-groups. Slides
CERMICS, ENPC, PhD seminar, 2018, Crash course on ergodicity for Markov chains: doing probabilities like an analyst.
Analysis, 1st year undergraduates (differenciation, optimization...) , University Paris Dauphine, 2017.
Tutorship for analysis, 1st year at Ecole des Ponts (Sobolev spaces, distributions, Dirichlet problem...), 2017.
Project of 4 students at Ecole des Ponts, Branching algorithms for computing eigenvalues of operators, 2018.
Mentoring of a graduate-level six month internship, Stein Variational Gradient Descent algorithm, 2018.
Co-organizer of the working group of young researchers in Statistical Physics and Interactions (J-PSI). See the web page of the group here.
Member (elected) of the comity of Ponts Alumni.