En poursuivant votre navigation sur ce site, vous acceptez le dépôt de cookies dans votre navigateur. (En savoir plus)
Portail > Offres > Offre UMR8538-HARBHA-006 - POSTDOC H/F

POSTDOC M/F

This offer is available in the following languages:
Français - Anglais

Date Limite Candidature : mardi 7 décembre 2021

Assurez-vous que votre profil candidat soit correctement renseigné avant de postuler. Les informations de votre profil complètent celles associées à chaque candidature. Afin d’augmenter votre visibilité sur notre Portail Emploi et ainsi permettre aux recruteurs de consulter votre profil candidat, vous avez la possibilité de déposer votre CV dans notre CVThèque en un clic !

Faites connaître cette offre !

General information

Reference : UMR8538-HARBHA-006
Workplace : PARIS 05
Date of publication : Tuesday, November 16, 2021
Type of Contract : FTC Scientist
Contract Period : 24 months
Expected date of employment : 1 March 2022
Proportion of work : Full time
Remuneration : 3200 Euros per month
Desired level of education : PhD
Experience required : Indifferent

Missions

The objective here is to rely on simulations to back out physical characteristics of faults in Earth (in terms of friction, stresses, tractions, etc.) in a way that reduces computing costs. The models in WP1 and WP2 rely on a number of parameters that are not easy to constrain, and our objective is to use active learning techniques to vastly reduce the computational power needed to explore this parameter space, and find the best parameter sets. Active learning is a branch of machine learning that deals with building models when new data points are difficult to obtain (for example, they require running costly simulations or experiments). In our case, the algorithm tries to build a model linking the input parameters of the simulation to its output; when probing this model, it identifies the regions in the parameter space it is most unsure about, and decides to create new data points (in our example, run new simulations) to probe these regions. Once the model becomes strong enough, it will ultimately be used instead of the simulations to instantly produce physically accurate predictive energy release scenario using surrogate modeling.

Activities

1. Develop active learning algorithms to accelerate computations in WP's 1&2
2. Reproduce historical data by finding the best associated simulation
3. Refine for more recent but continuous data

Skills

Fracture Mechanics, Seismology, Computational Earthquake Source Mechanics

Work Context

The post doctoral candidate would be based at the Laboratoire de Géologie (UMR 8538) at Ecole Normale Supérieure. Founded in 1880, the LG ENS is a joint research unit between the CNRS and ENS-PSL. Built on a long tradition in Earth and Environmental Sciences, it houses research that covers a wide field - Geology, Geodesy, Geomorphology, Geodynamics, Marine Geophysics, Geomechanics, Hydrogeology, Mineralogy, Seismology and Tectonics - which makes it a privileged place for exchanges at thematic borders.

Constraints and risks

None

We talk about it on Twitter!