Learning Hierarchical and Topographic Dictionaries with Structured Sparsity - Inria - Institut national de recherche en sciences et technologies du numérique Access content directly
Conference Papers Year : 2011

Learning Hierarchical and Topographic Dictionaries with Structured Sparsity


Recent work in signal processing and statistics have focused on defining new regularization functions, which not only induce sparsity of the solution, but also take into account the structure of the problem. We present in this paper a class of convex penalties introduced in the machine learning community, which take the form of a sum of l_2 and l_infinity-norms over groups of variables. They extend the classical group-sparsity regularization in the sense that the groups possibly overlap, allowing more flexibility in the group design. We review efficient optimization methods to deal with the corresponding inverse problems, and their application to the problem of learning dictionaries of natural image patches: On the one hand, dictionary learning has indeed proven effective for various signal processing tasks. On the other hand, structured sparsity provides a natural framework for modeling dependencies between dictionary elements. We thus consider a structured sparse regularization to learn dictionaries embedded in a particular structure, for instance a tree or a two-dimensional grid. In the latter case, the results we obtain are similar to the dictionaries produced by topographic independent component analysis.
Fichier principal
Vignette du fichier
wavelets.pdf (567.47 Ko) Télécharger le fichier
Origin : Files produced by the author(s)

Dates and versions

inria-00633983 , version 1 (20-10-2011)



Julien Mairal, Rodolphe Jenatton, Guillaume Obozinski, Francis Bach. Learning Hierarchical and Topographic Dictionaries with Structured Sparsity. SPIE Wavelets and Sparsity XIV, Aug 2011, San Diego, United States. ⟨10.1117/12.893811⟩. ⟨inria-00633983⟩
165 View
255 Download



Gmail Facebook Twitter LinkedIn More