PhD Position F/M PhD - Robust few-shot learning for foundational model in CT imaging

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

Type de contrat : CDD

Niveau de diplôme exigé : Bac + 5 ou équivalent

Fonction : Doctorant

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 41 project teams, 27 of which operate jointly with Paris-Saclay University (15 teams) and the Institut Polytechnique de Paris (12 teams). Its activities occupy over 600 people, scientists and research and innovation support staff, including 44 different nationalities.

The centre also hosts the Institut DATAIA, dedicated to data sciences and their disciplinary and application interfaces.

Contexte et atouts du poste

The Greater Paris University Hospitals Data Warehouse (EDS AP-HP) contains multimodal clinical data (PMSI, imaging, biological, and clinical documents) for over 14 million patients. The ANR FM2AI projet proposes to leverage 50,000 real-world clinical 3D CT scans from this exceptional data resource, to deploy a novel foundation model for abdominal-pelvic CT Imaging. The approach is designed to generalize across multiple clinical applications involving abdominal CT images, by resorting to self-supervised learning techniques for training the
foundation model, and then exploiting it for a wide class of clinical queries thanks to the innovative
few-shot learning paradigm [1], while paying attention to robustness assessment.

In this context, we are seeking for a PhD candidate with an excellent background in AI and mathematics, to design robust few-shot learning methods to allow the on-site adaptation of the foundation model and generalization to specific diagnostic tasks, such as prediction and segmentation of CT images of all body regions, without requiring massive re-annotation efforts nor GPU resources. The work will build upon the expertise of the OPIS team on few-short learning [2,3,4,5]. 

[1] E. Pachetti, S. Colantonio, A systematic review of few-shot learning in medical imaging, Art. Int. Med., 2024.

[2] S. Martin, M. Boudiaf, E. Chouzenoux, J.-C. Pesquet, et al., Towards practical few-shot query sets:
Transductive minimum description length inference, Proc. the Int. Conf. on Neu. Inf. Proc. Sys. (NeurIPS), 2022.

[3] S. Martin, Y. Huang, F. Shakeri, J.-C. Pesquet, I. Ben Ayed, Transductive zero-shot and few-shot CLIP, IEEE
/ CVF Computer Vision and Pattern Recognition Conference (CVPR), 2024.

[4] L. Zhou, F. Shakeri, A. Sadraoui, M. Kaaniche, J.-C. Pesquet, I. Ben Ayed, UNEM: UNrolled Generalized EM
for Transductive Few-Shot Learning, IEEE/CVF Conf. on Comp. Vision and Patt. Recognition (CVPR), 2025.

[5] M. Vu, E. Chouzenoux, J.-C. Pesquet, I. Ben-Ayed. Aggregated f-average Neural Network applied to Few-
Shot Class Incremental Learning, vol. 237, pp. 110054, Signal Processing, 2025.

Mission confiée

Missions: Develop new few-shot learning techniques for CT image classification ; Develop new model for few-shot tumor segmentation; Analyze robustness and generalization capabilities of the models ; Validation on public datasets and EDS-APHP datasets. 

Environment: The phd student will be supervised by Emilie Chouzenoux (Head of OPIS team, Inria Saclay), and will interact regularly with the members of the ANR FM2AI consortium. The 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.

Starting date is flexible, from the 1st Oct. 2026.

Principales activités

Main activities :

Programming in Python/PyTorch environment

Bibliographical study

Deep learning architecture design/training/testing

Mathematical optimization / convergence analysis

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

2300€ gross/month