Fast Gaussian Pairwise Constrained Spectral Clustering

David Chatel 1, * Pascal Denis 1 Marc Tommasi 1
* Auteur correspondant
1 MAGNET - Machine Learning in Information Networks
LIFL - Laboratoire d'Informatique Fondamentale de Lille, Inria Lille - Nord Europe
Abstract : We consider the problem of spectral clustering with partial supervision in the form of must-link and cannot-link constraints. Such pairwise constraints are common in problems like coreference resolution in natural language processing. The approach developed in this paper is to learn a new representation space for the data together with a dis-tance in this new space. The representation space is obtained through a constraint-driven linear transformation of a spectral embedding of the data. Constraints are expressed with a Gaussian function that locally reweights the similarities in the projected space. A global, non-convex optimization objective is then derived and the model is learned via gradi-ent descent techniques. Our algorithm is evaluated on standard datasets and compared with state of the art algorithms, like [14,18,31]. Results on these datasets, as well on the CoNLL-2012 coreference resolution shared task dataset, show that our algorithm significantly outperforms related approaches and is also much more scalable.
Type de document :
Communication dans un congrès
ECML/PKDD - 7th European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases, Sep 2014, Nancy, France. pp.242 - 257, 2014, 〈10.1007/978-3-662-44848-9_16〉
Liste complète des métadonnées

https://hal.inria.fr/hal-01017269
Contributeur : Team Magnet <>
Soumis le : vendredi 18 juillet 2014 - 15:35:43
Dernière modification le : jeudi 11 janvier 2018 - 06:25:27
Document(s) archivé(s) le : jeudi 20 novembre 2014 - 16:42:24

Fichier

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

Identifiants

Citation

David Chatel, Pascal Denis, Marc Tommasi. Fast Gaussian Pairwise Constrained Spectral Clustering. ECML/PKDD - 7th European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases, Sep 2014, Nancy, France. pp.242 - 257, 2014, 〈10.1007/978-3-662-44848-9_16〉. 〈hal-01017269〉

Partager

Métriques

Consultations de la notice

673

Téléchargements de fichiers

297