Efficient Learning of Sparse Conditional Random Fields for Supervised Sequence Labelling - Archive ouverte HAL Accéder directement au contenu
Article Dans Une Revue IEEE J. Sel. Topics Signal Process Année : 2010

Efficient Learning of Sparse Conditional Random Fields for Supervised Sequence Labelling

Résumé

Conditional Random Fields (CRFs) constitute a popular and efficient approach for supervised sequence labelling. CRFs can cope with large description spaces and can integrate some form of structural dependency between labels. In this contribution, we address the issue of efficient feature selection for CRFs based on imposing sparsity through an L1 penalty. We first show how sparsity of the parameter set can be exploited to significantly speed up training and labelling. We then introduce coordinate descent parameter update schemes for CRFs with L1 regularization. We finally provide some empirical comparisons of the proposed approach with state-of-the-art CRF training strategies. In particular, it is shown that the proposed approach is able to take profit of the sparsity to speed up processing and hence potentially handle larger dimensional models.
Fichier principal
Vignette du fichier
SLCY09.pdf (265.69 Ko) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)
Loading...

Dates et versions

hal-00414107 , version 1 (07-09-2009)
hal-00414107 , version 2 (03-01-2010)

Identifiants

Citer

Nataliya Sokolovska, Thomas Lavergne, Olivier Cappé, François Yvon. Efficient Learning of Sparse Conditional Random Fields for Supervised Sequence Labelling. IEEE J. Sel. Topics Signal Process, 2010, 4 (6), pp.953-964. ⟨10.1109/JSTSP.2010.2076150⟩. ⟨hal-00414107v2⟩
101 Consultations
452 Téléchargements

Altmetric

Partager

Gmail Facebook X LinkedIn More