Nicolas Verzelen
About me
I work as a research scientist (aka chargé de recherche, hdr) in Statistics and Machine Learning within INRAE. I am mostly interested in the analysis of popular algorithms for modern data science with a particular emphasis on methods for agroecological data.
Research
My current research interests include
unsupervised learning (clustering, ranking)
high-dimensional minimax theory and computational trade-offs.
active learning
applications to seed exchange networks
applications to agroecology
Duties
Some recent publications
B. Even, C. Giraud, and N. Verzelen. Computation-information gap in high-dimensional clustering. Accepted in COLT 2024 [ArXiv]
J. Mariel, I. Sanchez, N. Verzelen, F. Massol, S. Carrière, and V. Labeyrie The role of farmers’ networks in sourcing planting material and information in a context of agroforestry transition in Madagascar. Agricultural Systems, 217, 2024 [Article]
E. Pilliat, A. Carpentier, and N. Verzelen. Optimal rates for ranking a permuted isotonic matrix in polynomial time. Symposium on Discrete Algorithms (SODA), 3236–3273, 2024. [Article] [ArXiv]
E. Saad, A. Carpentier, and N. Verzelen. Active Ranking of Experts Based on their Performances in Many Tasks. ICML (oral presentation), 2023. [Article] [ArXiv]
N. Verzelen, M. Fromont, M. Lerasle, and P. Reynaud-Bouret. Optimal Change-Point Detection and Localization. Annals of Statistics, 51(4):1586–1610, 2023. [Article] [ArXiv]
Full list of publications.
Supervision
Current Students:
Alumni:
Teaching
Currently:
|