Post-doctoral Research Visit F/m Multilevel - Lyon, France - Inria

Inria

Entreprise vérifiée

Lyon, France

il y a 3 semaines

Posté par:

Sophie Dupont

beBee Recruiter

Description

Le descriptif de l'offre ci-dessous est en Anglais_

Type de contrat :

CDD

Niveau de diplôme exigé :
Thèse ou équivalent

Fonction :
Post-Doctorant

A propos du centre ou de la direction fonctionnelle:
The Inria research centre in Lyon is the 9th Inria research centre, formally created in January 2022. It brings together approximately 300 people in 17 research teams and research support services.

Its staff are distributed in Villeurbanne, Lyon Gerland, and Saint-Etienne.

The Lyon centre is active in the fields of software, distributed and high-performance computing, embedded systems, quantum computing and privacy in the digital world, but also in digital health and computational biology.

Contexte et atouts du poste:

The postdoctoral research associate will be part of the OCKHAM team at ENS Lyon in Lyon, France and collaborate also with the SySiPh team.

ENS Lyon is a small-size research-driven university in France, that is consistently ranked among the best universities in France.

The applicant will benefit from a stimulating environment of research scientists in machine learning, signal processing and physics, with weekly seminars given by international experts.

Mission confiée:

The project will aim to adapt and extend multilevel optimization techniques to the following two settings:

-
Second order methods: Second order methods (i.e., approaches that use information coming from the second order derivatives), such as Newton method, which are known in smooth optimization to have a much faster convergence rate than first order ones [3]. They have been used also to accelerate the solution of inverse prob- lems in imaging [4] but not in a multilevel context. Indeed, such an formalism is not just a straightforward extension of the first order ML techniques and requires a new definition of function approximations at coarse levels that exploits also second order information in a non-smooth context.-
Stochastic multilevel methods: In order to combine acceleration with dimensionality reduction of the variables space and of the samples space, we aim at combining multilevel strategies with mini-batch strategies [5]. This requires the development of stochastic variants of multilevel algorithms, which has never been carried out, even

in the context of smooth optimization

References:

[1] G. Lauga, E. Riccietti, N. Pustelnik, and P.

Goncalves, "IML FISTA:
Inexact MuLti
level FISTA for Image Restoration", arXiv : , 202- [2] R. E. Carrillo, J. D. McEwen, and Y. Wiaux, "Sparsity averaging reweighted anal
ysis (SARA): a novel algorithm for radio-interferometric imaging", Monthly Notices of the

Royal Astronomical Society, vol. 426, no. 2, pp , 2012

- [3] J. Nocedal, and S.J. Wright, "Numerical optimization", Springer New York, 199- [4] S. Salzo, and S. Villa, "Convergence analysis of a proximal Gauss-Newton method."

Computational Optimization and Applications :

- [5] L. Bottou, and F. E. Curtis, and J. Nocedal. "Optimization methods for large-scale
- machine learning

" SIAM review :

Principales activités:

Main activities

Develop stochastic and Newton multilevel methods
Study the convergence of the proposed methods
Implement the methods in python
Test the methods on realistic problems
Compare with the stateoftheart

Compétences:

Technical skills and level required:

Languages:

Relational skills:

Other valued appreciated:

Avantages:

Subsidized meals
Partial reimbursement of public transport costs
Leave: 7 weeks of annual leave + 10 extra days off due to RTT (statutory reduction in working hours) + possibility of exceptional leave (sick children, moving home, etc.)
Possibility of teleworking (90 days / year) and flexible organization of working hours (except for intership)
Social, cultural and sports events and activities
Access to vocational training
Social security coverage under conditions

Rémunération:

2788 € gross salary / month

Informations générales:

-
Thème/Domaine: Optimisation, apprentissage et méthodes statistiques
Calcul Scientifique (BAP E)

-
Ville: Lyon

-
Centre Inria: Centre Inria de Lyon
-
Date de prise de fonction souhaitée:
-
Durée de contrat: 2 ans
-
Date limite pour postuler:

Consignes pour postuler:

Sécurité défense:

Ce poste est susceptible d'être affecté dans une zone à régime restrictif (ZRR), telle que définie dans le décret n° relatif à la protection du potentiel scientifique et technique de la nation (PPST).

L'autorisation d'accès à une zone est délivrée par le chef d'établissement, après avis ministériel favorable, tel que défini dans l'arrêté du 03 juillet 2012, relatif à la PPST.

Un avis ministériel défavorable pour un poste affecté dans une ZRR aurait pour conséquence l'annulation du recrutement.

**Polit