• Refine Query
  • Source
  • Publication year
  • to
  • Language
  • 2
  • Tagged with
  • 2
  • 2
  • 2
  • 2
  • 2
  • 2
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • About
  • The Global ETD Search service is a free service for researchers to find electronic theses and dissertations. This service is provided by the Networked Digital Library of Theses and Dissertations.
    Our metadata is collected from universities around the world. If you manage a university/consortium/country archive and want to be added, details can be found on the NDLTD website.
1

Structured sparsity-inducing norms : statistical and algorithmic properties with applications to neuroimaging

Jenatton, Rodolphe 24 November 2011 (has links) (PDF)
Numerous fields of applied sciences and industries have been recently witnessing a process of digitisation. This trend has come with an increase in the amount digital data whose processing becomes a challenging task. In this context, parsimony, also known as sparsity, has emerged as a key concept in machine learning and signal processing. It is indeed appealing to exploit data only via a reduced number of parameters. This thesis focuses on a particular and more recent form of sparsity, referred to as structured sparsity. As its name indicates, we shall consider situations where we are not only interested in sparsity, but where some structural prior knowledge is also available. The goal of this thesis is to analyze the concept of structured sparsity, based on statistical, algorithmic and applied considerations. To begin with, we introduce a family of structured sparsity-inducing norms whose statistical aspects are closely studied. In particular, we show what type of prior knowledge they correspond to. We then turn to sparse structured dictionary learning, where we use the previous norms within the framework of matrix factorization. From an optimization viewpoint, we derive several efficient and scalable algorithmic tools, such as working-set strategies and proximal-gradient techniques. With these methods in place, we illustrate on numerous real-world applications from various fields, when and why structured sparsity is useful. This includes, for instance, restoration tasks in image processing, the modelling of text documents as hierarchy of topics, the inter-subject prediction of sizes of objects from fMRI signals, and background-subtraction problems in computer vision.
2

Structured sparsity-inducing norms : statistical and algorithmic properties with applications to neuroimaging / Normes parcimonieuses structurées : propriétés statistiques et algorithmiques avec applications à l’imagerie cérébrale

Jenatton, Rodolphe 24 November 2011 (has links)
De nombreux domaines issus de l’industrie et des sciences appliquées ont été les témoins d’une révolution numérique. Cette dernière s’est accompagnée d’une croissance du volume des données, dont le traitement est devenu un défi technique. Dans ce contexte, la parcimonie est apparue comme un concept central en apprentissage statistique. Il est en effet naturel de vouloir exploiter les données disponibles via un nombre réduit de paramètres. Cette thèse se concentre sur une forme particulière et plus récente de parcimonie, nommée parcimonie structurée. Comme son nom l’indique, nous considérerons des situations où, au delà de la seule parcimonie, nous aurons également à disposition des connaissances a priori relatives à des propriétés structurelles du problème. L’objectif de cette thèse est d'analyser le concept de parcimonie structurée, en se basant sur des considérations statistiques, algorithmiques et appliquées. Nous commencerons par introduire une famille de normes structurées parcimonieuses dont les aspects statistiques sont étudiées en détail. Nous considérerons ensuite l’apprentissage de dictionnaires, où nous exploiterons les normes introduites précédemment dans un cadre de factorisation de matrices. Différents outils algorithmiques efficaces, tels que des méthodes proximales, seront alors proposés. Grâce à ces outils, nous illustrerons sur de nombreuses applications pourquoi la parcimonie structurée peut être bénéfique. Ces exemples contiennent des tâches de restauration en traitement de l’image, la modélisation hiérarchique de documents textuels, ou encore la prédiction de la taille d’objets à partir de signaux d’imagerie par résonance magnétique fonctionnelle. / Numerous fields of applied sciences and industries have been recently witnessing a process of digitisation. This trend has come with an increase in the amount digital data whose processing becomes a challenging task. In this context, parsimony, also known as sparsity, has emerged as a key concept in machine learning and signal processing. It is indeed appealing to exploit data only via a reduced number of parameters. This thesis focuses on a particular and more recent form of sparsity, referred to as structured sparsity. As its name indicates, we shall consider situations where we are not only interested in sparsity, but where some structural prior knowledge is also available. The goal of this thesis is to analyze the concept of structured sparsity, based on statistical, algorithmic and applied considerations. To begin with, we introduce a family of structured sparsity-inducing norms whose statistical aspects are closely studied. In particular, we show what type of prior knowledge they correspond to. We then turn to sparse structured dictionary learning, where we use the previous norms within the framework of matrix factorization. From an optimization viewpoint, we derive several efficient and scalable algorithmic tools, such as working-set strategies and proximal-gradient techniques. With these methods in place, we illustrate on numerous real-world applications from various fields, when and why structured sparsity is useful. This includes, for instance, restoration tasks in image processing, the modelling of text documents as hierarchy of topics, the inter-subject prediction of sizes of objects from fMRI signals, and background-subtraction problems in computer vision.

Page generated in 0.0873 seconds