A Distributed Frank-Wolfe Framework for ...
Type de document :
Rapport de recherche: Autre communication scientifique (congrès sans actes - poster - séminaire...)
Titre :
A Distributed Frank-Wolfe Framework for Learning Low-Rank Matrices with the Trace Norm
Auteur(s) :
Zheng, Wenjie [Auteur]
Machine Learning and Information Access [MLIA]
Bellet, Aurelien [Auteur]
Machine Learning in Information Networks [MAGNET]
Gallinari, Patrick [Auteur]
Machine Learning and Information Access [MLIA]
Machine Learning and Information Access [MLIA]
Bellet, Aurelien [Auteur]
Machine Learning in Information Networks [MAGNET]
Gallinari, Patrick [Auteur]
Machine Learning and Information Access [MLIA]
Institution :
Inria Lille
Date de publication :
2017
Mot(s)-clé(s) en anglais :
Frank–Wolfe algorithm
Low-rank learning
Distributed optimization
Trace norm
Multi-task learning
Multinomial logistic regression
Low-rank learning
Distributed optimization
Trace norm
Multi-task learning
Multinomial logistic regression
Discipline(s) HAL :
Informatique [cs]/Apprentissage [cs.LG]
Statistiques [stat]/Machine Learning [stat.ML]
Informatique [cs]/Calcul parallèle, distribué et partagé [cs.DC]
Statistiques [stat]/Machine Learning [stat.ML]
Informatique [cs]/Calcul parallèle, distribué et partagé [cs.DC]
Résumé en anglais : [en]
We consider the problem of learning a high-dimensional but low-rank matrix from a large-scale dataset distributed over several machines, where low-rankness is enforced by a convex trace norm constraint. We propose DFW-Trace, ...
Lire la suite >We consider the problem of learning a high-dimensional but low-rank matrix from a large-scale dataset distributed over several machines, where low-rankness is enforced by a convex trace norm constraint. We propose DFW-Trace, a distributed Frank-Wolfe algorithm which leverages the low-rank structure of its updates to achieve efficiency in time, memory and communication usage. The step at the heart of DFW-Trace is solved approximately using a distributed version of the power method. We provide a theoretical analysis of the convergence of DFW-Trace, showing that we can ensure sublinear convergence in expectation to an optimal solution with few power iterations per epoch. We implement DFW-Trace in the Apache Spark distributed programming framework and validate the usefulness of our approach on synthetic and real data, including the ImageNet dataset with high-dimensional features extracted from a deep neural network.Lire moins >
Lire la suite >We consider the problem of learning a high-dimensional but low-rank matrix from a large-scale dataset distributed over several machines, where low-rankness is enforced by a convex trace norm constraint. We propose DFW-Trace, a distributed Frank-Wolfe algorithm which leverages the low-rank structure of its updates to achieve efficiency in time, memory and communication usage. The step at the heart of DFW-Trace is solved approximately using a distributed version of the power method. We provide a theoretical analysis of the convergence of DFW-Trace, showing that we can ensure sublinear convergence in expectation to an optimal solution with few power iterations per epoch. We implement DFW-Trace in the Apache Spark distributed programming framework and validate the usefulness of our approach on synthetic and real data, including the ImageNet dataset with high-dimensional features extracted from a deep neural network.Lire moins >
Langue :
Anglais
Collections :
Source :
Fichiers
- https://hal.inria.fr/hal-01672066/document
- Accès libre
- Accéder au document
- http://arxiv.org/pdf/1712.07495
- Accès libre
- Accéder au document
- https://hal.inria.fr/hal-01672066/document
- Accès libre
- Accéder au document
- https://hal.inria.fr/hal-01672066/document
- Accès libre
- Accéder au document
- document
- Accès libre
- Accéder au document
- main.pdf
- Accès libre
- Accéder au document
- 1712.07495
- Accès libre
- Accéder au document