Greedy Sparsity-Constrained Optimization

被引:0
|
作者
Bahmani, Sohail [1 ]
Raj, Bhiksha [2 ]
Boufounos, Petros T. [3 ]
机构
[1] Carnegie Mellon Univ, Dept Elect & Comp Engn, Pittsburgh, PA 15213 USA
[2] Carnegie Mellon Univ, Language Technol Inst, Pittsburgh, PA 15213 USA
[3] Mitsubishi Elect Res Labs, Cambridge, MA 02139 USA
关键词
sparsity; optimization; compressed sensing; greedy algorithm; GENERALIZED LINEAR-MODELS; SIGNAL RECOVERY; PURSUIT; L(1);
D O I
暂无
中图分类号
TP [自动化技术、计算机技术];
学科分类号
0812 ;
摘要
Sparsity-constrained optimization has wide applicability in machine learning, statistics, and signal processing problems such as feature selection and Compressed Sensing. A vast body of work has studied the sparsity-constrained optimization from theoretical, algorithmic, and application aspects in the context of sparse estimation in linear models where the fidelity of the estimate is measured by the squared error. In contrast, relatively less effort has been made in the study of sparsity-constrained optimization in cases where nonlinear models are involved or the cost function is not quadratic. In this paper we propose a greedy algorithm, Gradient Support Pursuit (GraSP), to approximate sparse minima of cost functions of arbitrary form. Should a cost function have a Stable Restricted Hessian (SRH) or a Stable Restricted Linearization (SRL), both of which are introduced in this paper, our algorithm is guaranteed to produce a sparse vector within a bounded distance from the true sparse optimum. Our approach generalizes known results for quadratic cost functions that arise in sparse linear regression and Compressed Sensing. We also evaluate the performance of GraSP through numerical simulations on synthetic and real data, where the algorithm is employed for sparse logistic regression with and without l(2)-regularization.
引用
收藏
页码:807 / 841
页数:35
相关论文
共 50 条
  • [41] A bilinear formulation for vector sparsity optimization
    Peleg, Dori
    Meir, Ron
    SIGNAL PROCESSING, 2008, 88 (02) : 375 - 389
  • [42] Structured Sparsity through Convex Optimization
    Bach, Francis
    Jenatton, Rodolphe
    Mairal, Julien
    Obozinski, Guillaume
    STATISTICAL SCIENCE, 2012, 27 (04) : 450 - 468
  • [43] Online Linear Optimization with Sparsity Constraints
    Wang, Jun-Kun
    Lu, Chi-Jen
    Lin, Shou-De
    ALGORITHMIC LEARNING THEORY, VOL 98, 2019, 98
  • [44] A SPARSITY CONSTRAINED INVERSE PROBLEM TO LOCATE PEOPLE IN A NETWORK OF CAMERAS
    Alahi, Alexandre
    Boursier, Yannick
    Jacques, Laurent
    Vandergheynst, Pierre
    2009 16TH INTERNATIONAL CONFERENCE ON DIGITAL SIGNAL PROCESSING, VOLS 1 AND 2, 2009, : 22 - 28
  • [45] Risk-Constrained Microgrid Reconfiguration Using Group Sparsity
    Dall'Anese, Emiliano
    Giannakis, Georgios B.
    IEEE TRANSACTIONS ON SUSTAINABLE ENERGY, 2014, 5 (04) : 1415 - 1425
  • [46] Submodular optimization problems and greedy strategies: A survey
    Yajing Liu
    Edwin K. P. Chong
    Ali Pezeshki
    Zhenliang Zhang
    Discrete Event Dynamic Systems, 2020, 30 : 381 - 412
  • [47] A GREEDY GENETIC ALGORITHM FOR UNCONSTRAINED GLOBAL OPTIMIZATION
    ZHAO Xinchao(Key Laboratory of Mathematics Mechanization
    Journal of Systems Science & Complexity, 2005, (01) : 102 - 110
  • [48] Submodular optimization problems and greedy strategies: A survey
    Liu, Yajing
    Chong, Edwin K. P.
    Pezeshki, Ali
    Zhang, Zhenliang
    DISCRETE EVENT DYNAMIC SYSTEMS-THEORY AND APPLICATIONS, 2020, 30 (03): : 381 - 412
  • [49] Sampling Optimization for Printer Characterization by Greedy Search
    Morovic, Jan
    Arnabat, Jordi
    Richard, Yvan
    Albarran, Angel
    IEEE TRANSACTIONS ON IMAGE PROCESSING, 2010, 19 (10) : 2705 - 2711
  • [50] A GREEDY FORWARD-BACKWARD ALGORITHM FOR ATOMIC NORM CONSTRAINED MINIMIZATION
    Rao, Nikhil
    Shah, Parikshit
    Wright, Stephen
    Nowak, Robert
    2013 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH AND SIGNAL PROCESSING (ICASSP), 2013, : 5885 - 5889