Frumam Etage 2 - Toutes les réservations
Vendredi 13 décembre 2024

Heure espace buffet (couloir)
Réservations modérées
salle de réunion : LIS-2
Réservations modérées
salle des séminaires : SEMINAIRE-2ETG
Réservations modérées
08:00 - 08:30
08:30 - 09:00
09:00 - 09:30
09:30 - 10:00
10:00 - 10:30
10:30 - 11:00
11:00 - 11:30 Séminaire Rauzy
11:00 à 12:30
Séminaire
11:30 - 12:00
12:00 - 12:30
12:30 - 13:00
13:00 - 13:30
13:30 - 14:00
14:00 - 14:30 Séminaire signal-apprentissage : Marc Castella
14:00 à 15:30
Séminaire

Title ----- Unsupervised linear component analysis for (some) probability mixtures Abstract -------- We deal with a model where a set of observations is obtained by a linear superposition of unknown components called sources. The problem consists in recovering the sources without knowing the linear transform. We extend the well-known Independent Component Analysis (ICA) methodology. Instead of assuming independent source components, we assume that the source vector is a probability mixture of two distributions, only one of which satisfies the ICA assumptions. Concerning the other component of the mixture, two scenarios have been considered: - in case where a model is available, we show how the Iterative Conditional Estimation (ICE) can be applied - in case where the samples are concentrated in a restricted but unknown area, we propose to cluster them based on a data-driven distance. A theoretical grounding is provided through a link with the Christoffel function. The considered approaches are fully unsupervised and do not involve any learning based on a training set.
14:30 - 15:00
15:00 - 15:30
15:30 - 16:00
16:00 - 16:30
16:30 - 17:00
17:00 - 17:30
17:30 - 18:00
18:00 - 18:30
18:30 - 19:00
19:00 - 19:30
^ Haut de la page