Nicolas Jouvin

I am a researcher at INRAE in the MIA-Paris laboratory, working in statistics and machine learning. I mainly worked in computational statistics for inference of latent variable models. Recently, I’m also interested in physics-informed machine learning methods and their applications for inverse problems and generative modeling.

From 2020 to 2021, I was a postdoc at Ecole Centrale Lyon working with Yohann De Castro on learning mixture models with sparse regularisation on the space of measures. Prior to that, I completed my PhD on high-dimensional data and graph clustering with discrete latent variable models at Paris 1 Panthéon-Sorbonne University, under the supervision of Pierre Latouche, Charles Bouveyron and Alain Livartowski.

research interests:

Physics-informed machine learning (PINNs)
Latent variable models
Variational inference
Sparse regularisation

news

Oct 15, 2025	M2 Internship offer on parameter estimation in SDEs with PINNs.

selected publications

Preprint

Effective regions and kernels in continuous sparse regularisation, with application to sketched mixtures

Yohann De Castro, Rémi Gribonval, and Nicolas Jouvin

Jul 2025

preprint

Bib PDF

@unpublished{decastro:hal-05162134,
  title = {{Effective regions and kernels in continuous sparse regularisation, with application to sketched mixtures}},
  author = {De Castro, Yohann and Gribonval, R{\'e}mi and Jouvin, Nicolas},
  url = {https://hal.science/hal-05162134},
  year = {2025},
  month = jul,
}

Package

jinns: a JAX Library for Physics-Informed Neural Networks

Hugo Gangloff and Nicolas Jouvin

Dec 2024

https://mia_jinns.gitlab.io/jinns/

Bib PDF

@unpublished{gangloff:hal-04847347,
  title = {{jinns: a JAX Library for Physics-Informed Neural Networks}},
  author = {Gangloff, Hugo and Jouvin, Nicolas},
  url = {https://hal.inrae.fr/hal-04847347},
  year = {2024},
  month = dec,
}

Stat Comput

A Bayesian Fisher-EM algorithm for discriminative Gaussian subspace clustering

Nicolas Jouvin, Charles Bouveyron, and Pierre Latouche

Statistics and Computing, May 2021

The FisherEM package is available on CRAN, see https://github.com/nicolasJouvin/FisherEM for additional information

DOI Bib PDF

@article{jouvin:hal-03047930,
  title = {{A Bayesian Fisher-EM algorithm for discriminative Gaussian subspace clustering}},
  author = {Jouvin, Nicolas and Bouveyron, Charles and Latouche, Pierre},
  url = {https://hal.science/hal-03047930},
  journal = {{Statistics and Computing}},
  publisher = {{Springer Verlag (Germany)}},
  volume = {31},
  year = {2021},
  month = may,
  doi = {10.1007/s11222-021-10018-6},
}

ADAC

Hierarchical clustering with discrete latent variable models and the integrated classification likelihood

Etienne Come, Nicolas Jouvin, Pierre Latouche, and 1 more author

Advances in Data Analysis and Classification, May 2021

DOI Bib PDF

@article{come:hal-02530705,
  title = {{Hierarchical clustering with discrete latent variable models and the integrated classification likelihood}},
  author = {Come, Etienne and Jouvin, Nicolas and Latouche, Pierre and Bouveyron, Charles},
  url = {https://hal.science/hal-02530705},
  journal = {{Advances in Data Analysis and Classification}},
  publisher = {{Springer Verlag}},
  volume = {15},
  pages = {957--986},
  year = {2021},
  doi = {10.1007/s11634-021-00440-z},
}