beta-risk: a New Surrogate Risk for Learning from Weakly Labeled Data

Abstract : During the past few years, the machine learning community has paid attention to developing new methods for learning from weakly labeled data. This field covers different settings like semi-supervised learning, learning with label proportions, multi-instance learning, noise-tolerant learning, etc. This paper presents a generic framework to deal with these weakly labeled scenarios. We introduce the \betarisk as a generalized formulation of the standard empirical risk based on surrogate margin-based loss functions. This risk allows us to express the reliability on the labels and to derive different kinds of learning algorithms. We specifically focus on SVMs and propose a soft margin \betasvm algorithm which behaves better that the state of the art.
Type de document :
Communication dans un congrès
NIPS 2016, Dec 2016, Barcelona, Spain. Advances in Neural Information Processing Systems 29 (NIPS 2016)
Liste complète des métadonnées

Littérature citée [23 références]  Voir  Masquer  Télécharger

https://hal.archives-ouvertes.fr/hal-01359298
Contributeur : Valentina Zantedeschi <>
Soumis le : mardi 15 novembre 2016 - 15:57:48
Dernière modification le : jeudi 26 juillet 2018 - 01:11:07
Document(s) archivé(s) le : jeudi 16 mars 2017 - 13:25:06

Fichier

nips_site.pdf
Fichiers produits par l'(les) auteur(s)

Identifiants

  • HAL Id : hal-01359298, version 1

Citation

Valentina Zantedeschi, Rémi Emonet, Marc Sebban. beta-risk: a New Surrogate Risk for Learning from Weakly Labeled Data. NIPS 2016, Dec 2016, Barcelona, Spain. Advances in Neural Information Processing Systems 29 (NIPS 2016). 〈hal-01359298〉

Partager

Métriques

Consultations de la notice

447

Téléchargements de fichiers

179