Cet événement est organisé par la SFdS, plus précisément par les groupes spécialisés : Statistique bayésienne, Environnement et statistique, Statistique mathématique, et Fiabilité et incertitudes.
Dates : du 11 au 15 octobre 2027 au CIRM (Marseille, France).
Bayes at CIRM 2027 is a research school organized under the auspices of the Société Française de Statistique (SFdS), jointly coordinated by its specialized groups (Statistique bayésienne, Environnement et statistique, Statistique mathématique, and Fiabilité et incertitudes), to be held from October 11 to 15, 2027 at CIRM in Luminy, Marseille.
The development and analysis of robust Bayesian methods for high-dimensional statistical settings are fundamental in addressing modern real-world learning problems. These approaches are particularly attractive due to their inherent capability to quantify the uncertainty associated with any statistical estimation procedure by computing or approximating posterior distributions.
The school is centered around two main themes:
- Non-parametric Bayesian statistics, which includes methods such as Gaussian Processes and Dirichlet Process Mixtures, offering flexible tools for modeling complex, high-dimensional data without assuming a fixed parametric form.
- Stochastic Partial Differential Equations (SPDEs) for spatial statistics, providing a principled framework for modeling spatial and spatiotemporal processes with inherent randomness and complex dependencies.
The event will focus on both theoretical advancements and practical applications of these methods, featuring discussions on the latest developments in non-parametric Bayesian approaches, scalable inference algorithms, and innovations in SPDEs including efficient numerical methods and applications to real-world spatial data.
