WP3: High-dimensional causal inference
| Teams
Duration: 3 years
Budget: €190,000
Scientific Coordinator:
Rodolphe Thiébaut, Bordeaux Population Health Research Center, UMR1219, Bordeaux
Partners:
Quentin Clairon, Aurore Li, Cécilia Samieri, Ima Bernada, Bordeaux Population Health Research Center, UMR1219, Bordeaux
Grégory Nuel, Laboratory of Probability, Statistics and Modeling Pierre et Marie Curie Campus, UMR8001
| Context
In public health research, establishing causal relationships between environmental exposures and health outcomes is essential for designing effective interventions. Exposome studies encompass a wide range of biological, chemical, social, and physical factors encountered throughout life. This complexity presents a unique challenge for statistical analysis and causal inference.
| Description
Work Package 3 (WP3) focuses on the challenge of causal inference in high-dimensional settings. The objectives of WP1 and WP2 require establishing causal links between the exposome and outcomes of interest, such as respiratory health. The exposome includes various types of measurements from very different processes (biological, social, etc.), which means the impulse program faces the challenge of causal inference in a high-dimensional context.
| Objectives
The objective is to propose projects on causal inference in highly dimensional frameworks in order to accelerate research and the application of relevant methods for environmental epidemiology.
A PhD is included in this Work Package, aiming to develop inference methods based on deep learning for models, with a particular focus on mechanistic models based on ordinary differential equations (ODEs). These methods are designed to analyze repeated measures from observational or experimental studies (e.g., in vitro or in vivo).
Programme d'impulsion Exposome 