IHU Prometheus G561

Université Sorbonne Paris Nord; France; Sur place; EUR 25 000 - 35 000; Plein temps

Résumé du poste

A renowned university in France is offering a PhD position focused on designing a multimodal learning framework for early sepsis diagnosis. Candidates will work on developing innovative methods leveraging deep learning to enhance clinical relevance, emphasizing real-world application through robust validation on large healthcare datasets. The ideal candidate should possess a strong background in AI and multitask learning.

Qualifications

• Strong knowledge of multitask learning and transformer-based architectures.
• Experience with large-scale datasets and clinical data handling.
• Ability to develop methods for gradient-based optimization.

Responsabilités

• Design a multimodal and temporal multitask learning framework for sepsis diagnosis.
• Develop methods to manage task interference in machine learning.
• Validate the model on large-scale healthcare datasets.

Connaissances

Deep learning

AI for Health

Formation

PhD

Description du poste

Organisation/Company Université Sorbonne Paris Nord Research Field Computer science » Informatics Researcher Profile Recognised Researcher (R2) Leading Researcher (R4) First Stage Researcher (R1) Established Researcher (R3) Country France Application Deadline 1 Dec 2025 - 00:00 (UTC) Type of Contract Temporary Job Status Full-time Offer Starting Date 1 Jan 2026 Is the job funded through the EU Research Framework Programme? Not funded by a EU programme Is the Job related to staff position within a Research Infrastructure? No

Offer Description

PHD

This PhD project aims to design a multimodal and temporal multitask learning framework for early sepsis diagnosis and risk prediction. Sepsis is a life-threatening condition caused by a dysregulated immune response to infection, making early detection essential yet highly challenging due to heterogeneous symptoms and complex multimodal data (vitals, lab tests, imaging, and clinical text). The proposed research focuses on multitask learning (MTL), where several interrelated clinical tasks—such as risk prediction, severity classification, and biomarker identification—are learned simultaneously within a shared model. Building on transformer-based architectures, the work will develop methods to handle task interference and conflicting objectives through gradient-based bi-level optimization and task decomposition via representation learning. The model will also integrate temporal reasoning to capture sepsis progression and emphasize interpretability to ensure clinical relevance. Finally, it will be validated on large-scale datasets (MIMIC-VI and IHU cohorts) to evaluate robustness, generalization, and potential deployment in real-world healthcare systems.

Laboratory: Laboratoire d'Informatique de Paris Nord

Field of research: Deep learning , AI for Health

Skills:

Knowledge :