Master Internship - Stability Analysis of Deep Unrolled Networks

Télécharger l'offre au format PDF
Le descriptif de l’offre ci-dessous est en Anglais

Type de contrat : Stage

Niveau de diplôme exigé : Bac + 5 ou équivalent

Fonction : Stagiaire de la recherche

Niveau d'expérience souhaité : Jeune diplômé

A propos du centre ou de la direction fonctionnelle

The Inria Saclay-Île-de-France Research Centre was established in 2008. It has developed as part of the Saclay site in partnership with Paris-Saclay University and with the Institut Polytechnique de Paris .

The centre has 40 project teams , 32 of which operate jointly with Paris-Saclay University and the Institut Polytechnique de Paris; Its activities occupy over 600 people, scientists and research and innovation support staff, including 44 different nationalities.

 

Contexte et atouts du poste

Subject: 

The resolution of linear inverse problems, in signal/image processing, is traditionnaly made through the minimization of a penalized cost function. This last decade, a new class of approaches have emerged, based on the data-driven principle of deep unrolling. The iterations of the minimization algorithm are untied accross a finite number of layers, yielding a neural network whose parameters are learned by supervised training procedure. The performance of such techniques are excellent, on several datasets and applications. 

In this project, we investigate the robustness of such deep unrolled inversion methods to input perturbations. Robustness is a critical property given the known vulnerability of deep neural networks to adversarial perturbations The preliminary study [1] shows an innovant stability analysis based on fixed point theory. The goal of the internship is to go further, by studying another class of minimization algorithm, and by relaxing the theoretical assumptions of the theorems. 

 [1] E. Chouzenoux, C. Della Valle and J.-C. Pesquet. Stability Bounds for the Unfolded Forward-Backward Algorithm. Journal of Mathematical Imaging and Vision, vol. 67, no. 48, 2025.

Mission confiée

Missions: The recruited student will first propose a mathematical formulation of the problem, and second, a suitable fixed-point analysis to solve it. Numerical experiments on synthetic data will be implemented, to illustrate the validity of the theoretical results.

Environment: The intern will be supervised by Emilie Chouzenoux (Head of OPIS team, Inria Saclay). The intern student will join the Inria Saclay team OPIS (https://opis-inria.eu/). He/she will be located in the Centre de la Vision Numérique, in CentraleSupélec campus, Saclay, France. He/she will enjoy an international and creative environment where research seminars and reading groups take place very often. Informatic material expenses will be covered within the limits of the scale in force. The student will participate to regular collaborative meetings researchers from OPIS team (Emilie Chouzenoux, Jean-Christophe Pesquet).

Organization: The proposed offer is dedicated to internship of Master 2 / Engineering students. The starting/end dates are flexible, with a minimum duration of 5 months.

Principales activités

Main activities :

Bibliographical study

Mathematical problem formulation

Stability Analysis

Scientific meetings

Writing of scientific reports

Compétences

Languages : The candidate must be fluent in english and/or french languages.

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 and flexible organization of working hours
  • Professional equipment available (videoconferencing, loan of computer equipment, etc.)
  • Social, cultural and sports events and activities
  • Access to vocational training
  • Social security coverage

Rémunération

Gratification