A probabilistic incremental proximal ...
Document type :
Compte-rendu et recension critique d'ouvrage
DOI :
Title :
A probabilistic incremental proximal gradient method
Author(s) :
Akyildiz, Ömer Deniz [Auteur]
Universidad Carlos III de Madrid [Madrid] [UC3M]
Chouzenoux, Emilie [Auteur]
OPtimisation Imagerie et Santé [OPIS]
Organ Modeling through Extraction, Representation and Understanding of Medical Image Content [GALEN-POST]
Elvira, Víctor [Auteur]
Centre de Recherche en Informatique, Signal et Automatique de Lille - UMR 9189 [CRIStAL]
Institut TELECOM/TELECOM Lille1
Míguez, Joaquín [Auteur]
Universidad Carlos III de Madrid [Madrid] [UC3M]
Universidad Carlos III de Madrid [Madrid] [UC3M]
Chouzenoux, Emilie [Auteur]
OPtimisation Imagerie et Santé [OPIS]
Organ Modeling through Extraction, Representation and Understanding of Medical Image Content [GALEN-POST]
Elvira, Víctor [Auteur]
Centre de Recherche en Informatique, Signal et Automatique de Lille - UMR 9189 [CRIStAL]
Institut TELECOM/TELECOM Lille1
Míguez, Joaquín [Auteur]
Universidad Carlos III de Madrid [Madrid] [UC3M]
Journal title :
IEEE Signal Processing Letters
Pages :
1257-1261
Publisher :
Institute of Electrical and Electronics Engineers
Publication date :
2019-07
ISSN :
1070-9908
English keyword(s) :
Stochastic gradient
Extended Kalman filtering
Probabilistic optimization
Proximal algorithms
Extended Kalman filtering
Probabilistic optimization
Proximal algorithms
HAL domain(s) :
Sciences de l'ingénieur [physics]/Traitement du signal et de l'image [eess.SP]
English abstract : [en]
In this paper, we propose a probabilistic optimization method, named probabilistic incremental proximal gradient (PIPG) method, by developing a probabilistic interpretation of the incremental proximal gradient algorithm. ...
Show more >In this paper, we propose a probabilistic optimization method, named probabilistic incremental proximal gradient (PIPG) method, by developing a probabilistic interpretation of the incremental proximal gradient algorithm. We explicitly model the update rules of the incremental proximal gradient method and develop a systematic approach to propagate the uncertainty of the solution estimate over iterations. The PIPG algorithm takesthe form of Bayesian filtering updates for a state-space model constructed by using the cost function. Our framework makes it possible to utilize well-known exact or approximate Bayesian filters, such as Kalman or extended Kalman filters, to solve large scale regularized optimization problems.Show less >
Show more >In this paper, we propose a probabilistic optimization method, named probabilistic incremental proximal gradient (PIPG) method, by developing a probabilistic interpretation of the incremental proximal gradient algorithm. We explicitly model the update rules of the incremental proximal gradient method and develop a systematic approach to propagate the uncertainty of the solution estimate over iterations. The PIPG algorithm takesthe form of Bayesian filtering updates for a state-space model constructed by using the cost function. Our framework makes it possible to utilize well-known exact or approximate Bayesian filters, such as Kalman or extended Kalman filters, to solve large scale regularized optimization problems.Show less >
Language :
Anglais
Popular science :
Non
ANR Project :
Comment :
5 pages
Collections :
Source :
Files
- http://arxiv.org/pdf/1812.01655
- Open access
- Access the document
- 1812.01655
- Open access
- Access the document