The CREST Statistics group develops research and offers student training in theoretical and applied statistics.
Statistical methods play nowadays a key role in data science, machine learning and artificial intelligence. They have reached an unprecedented level of impact in recent years.
The group has got international recognition for contributions to various areas of statistical science including high-dimensional statistics, machine learning, bayesian analysis, statistical computation and simulation, nonparametric estimation, statistical analysis of networks, functional data analysis, dimension reduction, statistical optimal transport, game theory, quantum statistics, time series, and extreme value theory.
Our faculty serve on editorial boards of top-ranked journals in statistics such as Annals of Statistics, Biometrika, Journal of the American Statistical Association, Journal of the Royal Statistical Society, and on program committees of major conferences in machine learning (NeurIPS, ICML, COLT).
The Statistics group is composed of two units, one located at ENSAE (Palaiseau, Institut Polytechnique de Paris) and the other at ENSAI (Bruz). The ENSAE unit participates in the consortia Fondation Mathématique Jacques Hadamard, Hi!Paris – Paris Artificial Intelligence for Society and Business, Center for Data Science of the University Paris-Saclay. The ENSAI unit participates in the Graduate Schools DIGISPORT and CyberSchool.
The faculty teach in the Master programs “Data Science” of Institut Polytechnique de Paris, the Master “Mathématiques, vision, apprentissage” (MVA) of the University Paris-Saclay, and the Master “Statistics for Smart Data” of ENSAI.
In September 2026, a new program co-led by Vianney Perchet (CREST, ENSAE Paris) aiming to use AI to overcome some of the current limitations encountered in financial markets and quantitative investment in order to solve complexe problems will open. This program will complement École Polytechnique’s already extensive AI training offering in its MSc&T programs, which include the MSC&T “Trustworthy and Responsible AI (TRAI)” and the MSc&T “Data Science and AI for Business,” in partnership with HEC, ranked second in the world in the QS2026 in Business Analytics ranking, and the MSc&T “Visual and Creative Artificial Intelligence (Ai-VIC).”
Discover the AI for Markets and Quantitative Investment (MaQI) MSC&T program here.
statistics
Gradient-free stochastic optimization of derivatives under strong convexity
We consider the problem of minimizing the k-th order partial derivative f=∂kjg of an unknown function g along a fixed coordinate direction j, based on noisy queries of g. Assuming that g has Hölder ...
Arxiv, Statistics Theory, 2026
statistics
Optimal Estimation of Discrete Multiview Distributions under Heteroskedastic Multinomial Sampling
Multiview latent-variable models provide a fundamental framework for discrete data analysis, with applications to latent structure models, topic models, and mixtures of product distributions. In the d ...
Preprint, 2026
statistics
Generalized multi-view model: Adaptive density estimation under low-rank constraints
We study the problem of bivariate discrete or continuous probability density estimation under low-rank constraints. For discrete distributions, we assume that the two-dimensional array to estimate is ...
Journal of Machine Learning Research, v.26(236), 1–52, 2026