Research
My thesis :
My thesis was devoted to some contributions to variable selection in high dimensional and its uses in biology. More precisely, I am interested in well-founded and implementable methods based on data-driven penalty calibrations in Gaussian linear regression. The biological goal is to improve the regulation mechanism knowledge of Arabidospis thaliana genes from transcriptomic data.
Pre-Publications :
An overview of variable selection procedures using regularization paths in high-dimensional Gaussian linear regression. (submitted)
with M. Gallopin and M-L. Martin, 2024. [arXiv]
Trade-off between predictive performance and FDR control for high-dimensional Gaussian model selection. (submitted)
with M-L. Martin, 2023. [arXiv]
with the supplementary material available here
International talks :
A non-asymptotic control of a kernel two-sample test. - Statistics Team seminar, University of Cambridge, Cambridge
Some statistical methods of dimension reduction: from PCA to t-SNE. - Random Geometry conference – CIRM – January, 2023.
Penalties calibrations in Gaussian linear regression with a high-dimensional context. - StatMathAppli 2022 – Fréjus – September, 2022.
Talks :
A non-asymptotic control of a kernel two-sample test. - Rencontres Statistiques du CEREMADE, Université Paris-Dauphine, Paris
------------------------------------------------------- - Mathématiques appliquées seminar, Laboratoire de Mathématiques Jean Leray, Nantes
------------------------------------------------------- - IMAG Team seminar, Université de Montpellier, Montpellier
------------------------------------------------------- - LAGA Team seminar, Université Sorbonne Paris Nord, Villetaneuse
------------------------------------------------------- - MIA Paris-Saclay Team seminar, AgroParisTech, Palaiseau
------------------------------------------------------- - Stats-Optim Team seminar, Institut de Mathématiques de Toulouse, Toulouse
------------------------------------------------------- - DATA Team seminar, Laboratoire Jean Kuntzmann, Grenoble
Generalization of the slope heuristics method for variable selection in high-dimensional Gaussian linear regression. - Statistique, Probabilités, Optimisation et Contrôle team seminar – Institut de Mathématiques de Bourgorgne, Dijon – Avril, 2023.
Model selection and penalty calibration for high-dimensional Gaussian linear regression. - Probabilités, Statistique, Physique Mathématique team seminar – Institut Camille Jordan, Lyon – March, 2023.
Contributions to variable selection in high dimension and its uses in biology. - Hub Biocomputing Seminar - ENS de Lyon, Lyon
Contributions to variable selection in high dimension and its uses in biology. - Soutenance de thèse - LMO, Univ. Paris-Saclay – December, 2022.
Contributions to variable selection in high-dimension through penalty calibrations. - Working Group in Statistique et Génome Team – LaMME, Université d'Evry – November, 2022.
Penalties calibrations in Gaussian linear regression with a high-dimensional context. - Working Group BIGS – Université de Lorraine, Centre INRIA Nancy – July, 2022.
Slope heuristics in 2D for penalties calibration in Gaussian linear regression with a high-dimensional context. - JdS 2022 – Campus de la Doua, Univ. Claude Bernard, Lyon – June, 2022.
Around t-SNE, with Vincent Rivoirard (presentation of articles Hinton, G. E., & Roweis, S. (2002) and Van der Maaten, L., & Hinton, G. (2008) and Cai, T. T., & Ma, R. (2021) and Van Assel, H., Espinasse, T., Chiquet, J., & Picard, F. (2022)). - Pizzama team at LMO – Univ. Paris-Saclay – Apr, 2022.
Building an experiment network from statistical tools. - Symposium PhD/PD at IPS2 – Univ. Paris-Saclay – Mar, 2022.
Trade-off between Predictive Risk and False Discovery Rate in high-dimensional Gaussian linear regression. - MaIAGE – Jouy-en-Josas – Dec, 2021.
------------------------------------------------------- - Forum des jeunes mathématicien.ne.s – Besançon – Dec, 2021.
------------------------------------------------------- - MIA Paris – AgroParisTech – Nov, 2021.
From Graph Centrality to Data Depth, with Vincent Rivoirard (presentation of article Aamari, E., Arias-Castro, E., & Berenfeld, C. (2021)). - Pizzama team at LMO – Univ. Paris-Saclay – Nov, 2021.
A FDR control in a model selection procedure for a high-dimensional Gaussian linear regression. - JPS 2021 – Ile d'Oléron – Oct, 2021.
Slope heuristics for Gaussian linear regression in a high-dimensional context. - JdS 2021 – online – Jun, 2021.
Data-driven penalties in high-dimensional Gaussian linear regression for variable identification. - Celeste INRIA team-project – online – May, 2021.
Gaussian linear regression in a high-dimensional context. - Vulgarization seminar – LMO, Univ. Paris-Saclay – May, 2021.
High-dimensional linear regression applied to gene interactions network inference. - Labmeeting at IPS2 – online – Dec, 2020.
Review and comparison of some methods of parameters selection to identify gene interactions in Arabidopsis thaliana. - NetBio – IPS2 – Oct, 2019.
Other :
Editorial activities: Referee for the journal Stochastic Environmental Research and Risk Assessment
Co-organization of activities of the French Statistical Society (SFdS)'s Young Statisticians group
Co-organization of the seminar for master students from Statistics and Machine learning and Artificial intelligence at Univ. Paris-Saclay
2021-2022: with Wojciech Reise, Matthieu Lerasle & Gilles Blanchard
2020-2021: with Vincent Divol, Matthieu Lerasle & Gilles Blanchard
Co-supervision of a master project (2020-2021) with Marie-Laure Martin