You are here : Version anglaise > Fellows > Fellows 2025-2026

Laurent
JACQUES

Computing - Belgium

Contact details

Email.
laurent.jacques@uclouvain.be
Personal Website.
https://laurentjacques.gitlab.io

Research topics

SCIENTIFIC PROJECT

"Novel learning algorithms for imaging sciences: leveraging random dataset summaries for uncertainty quantification and equivariant imaging procedures"

 

In imaging sciences, one needs to estimate various statistical properties of an object of interest. When observing an object in magnetic resonance imaging or radio astronomy (an organ or a galaxy), one aims to determine whether its features are real or merely imaging artifacts, thereby quantifying confidence in a given observation. When analyzing a collection of audio recordings, an ensemble of images, or user interactions in a social network, these elements can be studied as being generated by a "random process" whose statistical properties explain the structure of the set: the existence of user communities or clusters of images. In this research project, we propose to extract such statistical properties using novel algorithms and mathematical tools that rely on specific random summaries of the datasets under consideration. To achieve this, we will leverage intrinsic geometric invariances (translation or rotation invariance) to demonstrably enhance the design of these algorithms.

 

Activities / Resume

BIOGRAPHY

Laurent Jacques holds a PhD in mathematical physics and is a professor in the Mathematical Engineering Department at the ICTEAM institute of UCLouvain, Belgium.