hal-00804592, version 2
A Note on k-support Norm Regularized Risk Minimization
Résumé : The k-support norm has been recently introduced to perform correlated sparsity regularization. Although Argyriou et al. only reported experiments using squared loss, here we apply it to several other commonly used settings resulting in novel machine learning algorithms with interesting and familiar limit cases. Source code for the algorithms described here is available.
- 1 : GALEN (INRIA Saclay - Ile de France)
- INRIA – Ecole Centrale Paris
- 2 : Centre de vision numérique (CVN)
- Ecole Centrale Paris
- Domaine : Informatique/Apprentissage
- Mots-clés : k-support norm – structured sparsity – regularization – least-squares – hinge loss – support vector machine – SVM – regularized logistic regression – AdaBoost – support vector regression – SVR
- Versions disponibles : v1 (26-03-2013) v2 (27-03-2013)
- hal-00804592, version 2
- http://hal.inria.fr/hal-00804592
- oai:hal.inria.fr:hal-00804592
- Contributeur : Matthew Blaschko
- Soumis le : Mercredi 27 Mars 2013, 17:21:28
- Dernière modification le : Mercredi 27 Mars 2013, 17:23:53
Documents associés
Exporter