UE Model exploration for approximation of complex, high-dimensional problems

Informations aux utilisateurs

Veuillez noter que vous consultez une page du catalogue de formation 2020-2021. Le recrutement est actuellement terminé pour les licences, licences professionnelles, masters, DUT et formations réglementées de santé. Pour consulter le catalogue des formations 2021-2022, cliquez sur le lien suivant.

Diplômes intégrant cet élément pédagogique :

Descriptif

Many industrial applications invole expensive computational codes which can take weeks or months to run. It is typical for weather prediction, in aerospace sector or in the civil engineering field. There is here an important (economic) challenge to reduce the computational cost by constructing a surrogate for the input-to-output relationship. Since only a few number of model runs is affordable, dedicated tools have been developed to exploit this type of “not-so-big” data sets. This lecture focuses on some of the most recent advances in that direction.

Pré-requis

Basic knowledge in probability and statistics

Compétences visées

The goal of this lecture is to address the difficult problem of approximating high-dimensional functions, meaning functions of a large number of parameters. The first part of the lecture is devoted to interpolation techniques via polynomial functions or via Gaussian processes. In the second part, we present two methods for reducing the dimension of the input parameters space, namely the Sliced Inverse Regression and the Ridge Function Recovery.

Bibliographie

References Springer Handbook on UQ, R. Ghanem, D. Higdon and H. Owhadi (Eds)

Informations complémentaires

Méthode d'enseignement : En présence
Lieu(x) : Grenoble - Domaine universitaire
Langue(s) : Anglais