Aller au contenu principal
Accueil

Main navigation

  • Actualités
    • Toutes les actualités
    • Séminaires - Soutenances
  • Présentation
    • CREATIS
    • Organigramme
    • Personnels
    • Effectifs
    • Contacts
    • Accès
  • Recherche
    • Equipes de recherche
    • Projets transversaux
    • Projets Structurants
    • Plateformes d'imagerie
    • Rapports d'activités
    • Notes d'information données
  • Contributions
    • Publications
    • Brevets
    • Logiciels
  • Formations
    • Implications dans les formations
    • Ecoles doctorales
  • Emplois et Stages
  • French French
  • English English
Search API form
User account menu
  • Account
    • Se connecter

Fil d'Ariane

  1. Accueil
  2. Selecting from an infinite set of features

Selecting from an infinite set of features

Ajouter à mon calendrier
Google Agenda Outlook Calendrier
We propose a principled framework for learning with infi nitely many features, situations that are usually induced by continuously parametrized feature extraction methods. Such cases occur for instance when considering Gabor-based features in computer vision problems or when dealing with Fourier features for kernel approximations. We cast the problem as the one of fi nding a fi nite subset of features that minimizes a regularized empirical risk. After having analyzed the optimality conditions of such a problem, we propose a simple algorithm which has the flavour of a column-generation technique. We also show that using Fourier-based features, it is possible to perform approximate infi nite kernel learning. Our experimental results on several datasets show the bene fits of the proposed approach in several situations including texture classi cation, pixel classification and large-scale kernelized problems (involving about 100 thousand examples).

Orateur

Rémi Flamary

Lieu

Salle de réunion (B502 4ème étage), Bâtiment Blaise Pascal

Date - horaires

mar 27/08/2013 - 17:00

Type d'évenement

Séminaire

Barre liens pratiques

  • Authentication
  • Intranet
  • Flux rss
  • Creatis sur Twitter
  • Webmail
Accueil

Footer menu

  • Contact
  • Accès
  • Newsletter
  • Mentions Légales