Nicolas Verzelen
 |
Research Scientist, |
About me
I work as a research scientist (aka chargé de recherche) in Statistics within INRAE. I am mostly interested in the analysis of popular algorithms for modern data science with a particular emphasis on methods for agroecology data.
Research
My current research interests include
clustering
network analysis
change-point detection
minimax theory
applications to seed exchange networks
applications to agroecology
Duties
I currently serve as an associate editor for the Annals of Statistics, Bernoulli, and as an editor in chief for ESAIM: P&S.
I am in charge of the Probability and Statistics team of MISTEA
Some recent publications
S. Thépaut and N. Verzelen. Optimal Estimation of Schatten Norms of a rectangular Matrix. [ArXiv]
C. Giraud, Y. Issartel, and N. Verzelen Localization in 1D non-parametric latent space models from pairwise affinities [ArXiv]
E. Pilliat, A. Carpentier, and N. Verzelen. Optimal multiple change-point detection for high-dimensional data. [ArXiv]
N. Verzelen, M. Fromont, M. Lerasle, and P. Reynaud-Bouret. Optimal Change-Point Detection and Localization. [ArXiv]
E. Roquain and N. Verzelen. On using empirical null distributions in Benjamini-Hochberg procedure. [ArXiv] [Vignette]
A. Carpentier and N. Verzelen. Adaptive estimation of the sparsity in the Gaussian vector model. Annals of Statistics, 47(1):93–126, 2019. [Article] [ArXiv]
C. Giraud and N. Verzelen. Partial recovery bounds for clustering with the relaxed K-means. Mathematical Statistics and Learning, 1(3):317–374, 2018. [Article] [ArXiv]
Supervision
Current PhD Students:
Emmanuel Pilliat co-supervised with Alexandra Carpentier and Joseph Salmon [2020–2023]
Alumni:
Yann Issartel [PhD 2017–2020], co-supervised with Christophe Giraud. Now Post-doc at CREST ENSAE.
Solène Thépaut [PhD 2016–2019] co-supervised with Christophe Giraud. Now at NukkAI.
Teaching
Currently:
Statistical Machine Learning [HMMA308], jointly with Joseph Salmon and Benjamin Charlier.