Professor of Applied Mathematics

Ecole Polytechnique


Inria XPOP


I am Professor at the Applied Math department of École polytechnique where I have been working since September 2013. My lab is the CMAP of École polytechnique. I work on problems involving data in signal processing, statistics and learning with a taste for applications. I am the cohead of the SIMPAS team (Signal IMage numerical ProbAbilities Statistical Learning). I am also a member of the Xpop Inria team. I am in charge of many training programs at Polytechnique: Data Science last year (PA), Data Science for Business MScT and Data Science continuing education programs Starter Program of Polytechnique Executive Education. I have been involved in the creation of the Datascience track of Institut Polytechnique de Paris.


  • Data Science
  • Statistics and Statistical Learning
  • Signal Processing
  • R and Python Data Science stack
  • Entrepreunership


  • HdR, 2013

    Université Paris Sud

  • PhD in Applied Mathematics, 2002

    Ecole Polytechnique

  • Master in Mathematics and Artifical Intelligence, 1997

    ENS Cachan


Selected publications

Last Preprints

For more details, see the research page or the comprehensive list of publications


Current students

Former students

  • Rémi Besson, PhD with S. Allassonière, 2016-2019, Rare Disease Diagnosis Optimization – now Post Doc at Université de Paris
  • Marco Brigham, Post Doc, 2017-2019, Morpheo project, EEG Analysis – now CSO at Moonoia
  • Esther Derman, PhD, 2016-2017, Model Based Histogram Clustering – moved to Israel, now PhD student in the Technion
  • Solenne Thivin, PhD, 2012-2015, CIFRE Thales Optronics, Detection on Complex Textures – now professor in high school
  • Lucie Montuelle, PhD, 2011-2014, Conditional Density Estimation – now Data Scientist at RTE
  • Joseph Salmon, PhD, 2007-2010, Estimator Aggregation – now full professor at Université de Montpellier

Recent Blog Posts

My most recent blog post on miscellaneous themes…

Data Science References

I’m involved in several Data Science programs in both initial and continuing education. They differ depending on the audiences but the backbones are the same. In order to enjoy such programs, a basic knowledge in Math, Computer Science and Programming is required.

ggwordcloud: a word cloud geom for ggplot2

ggwordcloud provides a word cloud text geom for ggplot2. The placement algorithm implemented in C++ is an hybrid between the one of wordcloud and the one of wordcloud2.js. The cloud can grow according to a shape and stay within a mask.

Visualization Examples

A ggplot2 tour of good and bad visualization examples

Supervised Classification: an Exploration with R

The twoClass dataset Statistical point of view Generative Modeling Logistic regression Kernel Methods Optimization point of view SVM NN Trees Model comparison Global comparison K-NN Comparison We will use the twoClass dataset from Applied Predictive Modeling, the book of M.

From Parallel Plot to Radar Plot

ggplot2 provides a very elegant way to describe graphics. In this example, we will use its grammar to show how the parallel plot and the radar plots are related.