DocumentCode :
59623
Title :
Simultaneously Structured Models With Application to Sparse and Low-Rank Matrices
Author :
Oymak, Samet ; Jalali, Amin ; Fazel, Maryam ; Eldar, Yonina C. ; Hassibi, Babak
Author_Institution :
Dept. of Electr. Eng. & Comput. Sci., Univ. of California at Berkeley, Berkeley, CA, USA
Volume :
61
Issue :
5
fYear :
2015
fDate :
May-15
Firstpage :
2886
Lastpage :
2908
Abstract :
Recovering structured models (e.g., sparse or group-sparse vectors, low-rank matrices) given a few linear observations have been well-studied recently. In various applications in signal processing and machine learning, the model of interest is structured in several ways, for example, a matrix that is simultaneously sparse and low rank. Often norms that promote the individual structures are known, and allow for recovery using an order-wise optimal number of measurements (e.g., 11 norm for sparsity, nuclear norm for matrix rank). Hence, it is reasonable to minimize a combination of such norms. We show that, surprisingly, using multiobjective optimization with these norms can do no better, orderwise, than exploiting only one of the structures, thus revealing a fundamental limitation in sample complexity. This result suggests that to fully exploit the multiple structures, we need an entirely new convex relaxation. Further, specializing our results to the case of sparse and low-rank matrices, we show that a nonconvex formulation recovers the model from very few measurements (on the order of the degrees of freedom), whereas the convex problem combining the 11 and nuclear norms requires many more measurements, illustrating a gap between the performance of the convex and nonconvex recovery problems. Our framework applies to arbitrary structure-inducing norms as well as to a wide range of measurement ensembles. This allows us to give sample complexity bounds for problems such as sparse phase retrieval and low-rank tensor completion.
Keywords :
compressed sensing; matrix algebra; optimisation; convex recovery problem; low-rank matrices; low-rank tensor completion; multiobjective optimization; nonconvex recovery problem; order-wise optimal number; recovering structured models; sample complexity; simultaneously structured model; sparse matrices; sparse phase retrieval; Complexity theory; Matrix decomposition; Nuclear measurements; Sparse matrices; Symmetric matrices; Tensile stress; Vectors; Compressed sensing; compressed sensing; convex relaxation; regularization; sample complexity;
fLanguage :
English
Journal_Title :
Information Theory, IEEE Transactions on
Publisher :
ieee
ISSN :
0018-9448
Type :
jour
DOI :
10.1109/TIT.2015.2401574
Filename :
7036090
Link To Document :
بازگشت