Guarantees for Approximate Incremental SVMs - Archive ouverte HAL Accéder directement au contenu
Communication Dans Un Congrès Année : 2010

Guarantees for Approximate Incremental SVMs

Résumé

Assume a teacher provides examples one by one. An approximate incremental SVM computes a sequence of classifiers that are close to the true SVM solutions computed on the successive incremental training sets. We show that simple algorithms can satisfy an averaged accuracy criterion with a computational cost that scales as well as the best SVM algorithms with the number of examples. Finally, we exhibit some experiments highlighting the benefits of joining fast incremental optimization and curriculum and active learning (Schon and Cohn, 2000; Bordes et al., 2005; Bengio et al., 2009).
Fichier principal
Vignette du fichier
usunier10aistats.pdf (354.8 Ko) Télécharger le fichier
Origine : Fichiers éditeurs autorisés sur une archive ouverte
Loading...

Dates et versions

hal-00750932 , version 1 (12-11-2012)

Identifiants

  • HAL Id : hal-00750932 , version 1

Citer

Nicolas Usunier, Antoine Bordes, Léon Bottou. Guarantees for Approximate Incremental SVMs. 13th International Conference on Artificial Intelligence and Statistics, May 2010, Chia Laguna Resort, Sardinia, Italy. pp.884-891. ⟨hal-00750932⟩
169 Consultations
73 Téléchargements

Partager

Gmail Facebook X LinkedIn More