%0 Journal Article
%T A Clustering Approach to Learn Sparsely-Used Overcomplete Dictionaries
%A Alekh Agarwal
%A Animashree Anandkumar
%A Praneeth Netrapalli
%J Mathematics
%D 2013
%I arXiv
%X We consider the problem of learning overcomplete dictionaries in the context of sparse coding, where each sample selects a sparse subset of dictionary elements. Our main result is a strategy to approximately recover the unknown dictionary using an efficient algorithm. Our algorithm is a clustering-style procedure, where each cluster is used to estimate a dictionary element. The resulting solution can often be further cleaned up to obtain a high accuracy estimate, and we provide one simple scenario where $\ell_1$-regularized regression can be used for such a second stage.
%U http://arxiv.org/abs/1309.1952v2