Descriptif :
The team's objective is to develop methodological research in signal processing, image processing, and data science with a focus on machine learning and AI. The work involves designing interpretable models that can faithfully represent the observed multidimensional data while maintaining reasonable complexity, typically linear with the number of dimensions. It also involves developing efficient algorithms, accompanied by guarantees, that allow for reconstruction, inversion, and decision-making directly from reduced complexity models. The general scope of these developments enables applications across two main areas: biology and digital health, and chemometrics and the processing of signals from physical measurements. This work is carried out in collaboration with other projects within the BioSiS department and with national and international academic or industrial partners.
Mots-clés : Learning, Inverse problems, Tensor decompositions, Low-rank models, Hyperspectral and polarimetric imaging, SparsityContact : Sebastian Miron
