Description du sujet de thèse

Deep Generative Models to Decipher the Dynamics of Cellular Interactions.
The PhD student will develop deep generative models to decode and simulate the dynamics of cellular interactions from videomicroscopy data. The main objective will be to design a system capable of generating synthetic video sequences that reflect cellular behaviors under different experimental conditions. This approach will help identify and quantify dynamic differences linked to treatments, cellular states, or specific biological contexts.
The candidate will contribute to model architecture design, validation protocol development, and collaborative experiments with biologists. They will also take part in disseminating the research results through scientific publications and conference presentations.
Activities:
- Design and training of conditional generative models for video sequences.
- Analysis of cell dynamics differences between experimental conditions.
- Validation on biological datasets from tumor models and embryology.
- Contribution to publications and scientific dissemination.

Contexte de travail

The project takes place in a multidisciplinary environment combining artificial intelligence, cellular imaging, and biology.
The student will work within IBENS, in a team specialized in image analysis and deep learning.

Contraintes et risques

Screen work.