BZPEER
More

Discover

About BZPEERSearchGraph

Research tools

HomeCollectionsResearchCollaborationPublish

Network

ExpertsCommunities

Opportunities

Listings

Account

Log inCreate account
Log inCreate account

Source author record

Naihua Xiu

Explore this author’s publications, collaborators and research topics. This catalog record has not been claimed by its author.

ResearcherUnclaimed source record
math.OCInformation Theorymath.ITMachine Learningmath.SPmath.STStatistics Theory

Catalog footprint

Research overview

9works
7topics
4close collaborators

Actions

Is this your work?

Log in to claim
Open graphBrowse papers

Published work

9 publications

preprint2021arXiv

Computing One-bit Compressive Sensing via Double-Sparsity Constrained Optimization

One-bit compressive sensing gains its popularity in signal processing and communications due to its low storage costs and low hardware complexity. However, it has been a challenging task to recover the signal only by exploiting the one-bit (the sign) information. In this paper, we appropriately formulate the one-bit compressive sensing into a double-sparsity constrained optimization problem. The first-order optimality conditions for this nonconvex and discontinuous problem are established via the newly introduced ττ-stationarity, based on which, a gradient projection subspace pursuit (\texttt{GPSP}) algorithm is developed. It is proven that \texttt{GPSP} can converge globally and terminate within finite steps. Numerical experiments have demonstrated its excellent performance in terms of a high order of accuracy with a fast computational speed.

Open access
preprint2016arXiv

Z-tensors and complementarity problems

Tensors are multidimensional analogs of matrices. In this paper, based on degree-theoretic ideas, we study homogeneous nonlinear complementarity problems induced by tensors. By specializing this to ZZ-tensors (which are tensors with non-positive off-diagonal entries), we describe various equivalent conditions for a ZZ-tensor to have the global solvability property. We show by an example that the global solvability need not imply unique solvability and provide a sufficient and easily checkable condition for unique solvability.

Open access
preprint2015arXiv

The Sparsest Solutions to ZZ-Tensor Complementarity Problems

Finding the sparsest solutions to a tensor complementarity problem is generally NP-hard due to the nonconvexity and noncontinuity of the involved ℓ0\ell_0 norm. In this paper, a special type of tensor complementarity problems with ZZ-tensors has been considered. Under some mild conditions, we show that to pursuit the sparsest solutions is equivalent to solving polynomial programming with a linear objective function. The involved conditions guarantee the desired exact relaxation and also allow to achieve a global optimal solution to the relaxed nonconvex polynomial programming problem. Particularly, in comparison to existing exact relaxation conditions, such as RIP-type ones, our proposed conditions are easy to verify.

Open access
preprint2014arXiv

Exact Recovery for Sparse Signal via Weighted l1l_1 Minimization

Numerical experiments in literature on compressed sensing have indicated that the reweighted l1l_1 minimization performs exceptionally well in recovering sparse signal. In this paper, we develop exact recovery conditions and algorithm for sparse signal via weighted l1l_1 minimization from the insight of the classical NSP (null space property) and RIC (restricted isometry constant) bound. We first introduce the concept of WNSP (weighted null space property) and reveal that it is a necessary and sufficient condition for exact recovery. We then prove that the RIC bound by weighted l1l_1 minimization is δak<a−1a−1+γ2δ_{ak}<\sqrt{\frac{a-1}{a-1+γ^2}}, where a>1a>1, 0<γ≤10<γ\leq1 is determined by an optimization problem over the null space. When γ<1γ< 1 this bound is greater than a−1a\sqrt{\frac{a-1}{a}} from l1l_1 minimization. In addition, we also establish the bound on δkδ_k and show that it can be larger than the sharp one 1/3 via l1l_1 minimization and also greater than 0.4343 via weighted l1l_1 minimization under some mild cases. Finally, we achieve a modified iterative reweighted l1l_1 minimization (MIRL1) algorithm based on our selection principle of weight, and the numerical experiments demonstrate that our algorithm behaves much better than l1l_1 minimization and iterative reweighted l1l_1 minimization (IRL1) algorithm.

Open access
preprint2014arXiv

Gradient Support Projection Algorithm for Affine Feasibility Problem with Sparsity and Nonnegativity

Let AA be a real M×NM \times N measurement matrix and b∈RMb\in \mathbb{R}^M be an observations vector. The affine feasibility problem with sparsity and nonnegativity (AFPSNAFP_{SN} for short) is to find a sparse and nonnegative vector x∈RNx\in \mathbb{R}^N with Ax=bAx=b if such xx exists. In this paper, we focus on establishment of optimization approach to solving the AFPSNAFP_{SN}. By discussing tangent cone and normal cone of sparse constraint, we give the first necessary optimality conditions, αα-Stability, T-Stability and N-Stability, and the second necessary and sufficient optimality conditions for the related minimization problems with the AFPSNAFP_{SN}. By adopting Armijo-type stepsize rule, we present a framework of gradient support projection algorithm for the AFPSNAFP_{SN} and prove its full convergence when matrix AA is ss-regular. By doing some numerical experiments, we show the excellent performance of the new algorithm for the AFPSNAFP_{SN} without and with noise.

Open access
preprint2014arXiv

Sparse and Low-Rank Covariance Matrices Estimation

This paper aims at achieving a simultaneously sparse and low-rank estimator from the semidefinite population covariance matrices. We first benefit from a convex optimization which develops l1l_1-norm penalty to encourage the sparsity and nuclear norm to favor the low-rank property. For the proposed estimator, we then prove that with large probability, the Frobenious norm of the estimation rate can be of order O(s(log⁡r)/n)O(\sqrt{s(\log{r})/n}) under a mild case, where ss and rr denote the number of sparse entries and the rank of the population covariance respectively, nn notes the sample capacity. Finally an efficient alternating direction method of multipliers with global convergence is proposed to tackle this problem, and meantime merits of the approach are also illustrated by practicing numerical simulations.

Open access
preprint2013arXiv

New RIC Bounds via l_q-minimization with 0<q<=1 in Compressed Sensing

The restricted isometry constants (RICs) play an important role in exact recovery theory of sparse signals via l_q(0<q<=1) relaxations in compressed sensing. Recently, Cai and Zhang[6] have achieved a sharp bound δ_tk<\sqrt{1-1/t} for t>=4/3 to guarantee the exact recovery of k sparse signals through the l_1 minimization. This paper aims to establish new RICs bounds via l_q(0<q<=1) relaxation. Based on a key inequality on l_q norm, we show that (i) the exact recovery can be succeeded via l_{1/2} and l_1 minimizations if δ_tk<\sqrt{1-1/t} for any t>1, (ii)several sufficient conditions can be derived, such as for any 0<q<1/2, δ_2k<0.5547 when k>=2, for any 1/2<q<1, δ_2k<0.6782 when k>=1, (iii) the bound on δ_k is given as well for any 0<q<=1, especially for q=1/2,1, we obtain δ_k<1/3 when k(>=2) is even or δ_k<0.3203 when k(>=3) is odd.

Open access
preprint2011arXiv

Sufficient Conditions for Low-rank Matrix Recovery, Translated from Sparse Signal Recovery

The low-rank matrix recovery (LMR) is a rank minimization problem subject to linear equality constraints, and it arises in many fields such as signal and image processing, statistics, computer vision, system identification and control. This class of optimization problems is N\N¶-hard and a popular approach replaces the rank function with the nuclear norm of the matrix variable. In this paper, we extend the concept of ss-goodness for a sensing matrix in sparse signal recovery (proposed by Juditsky and Nemirovski [Math Program, 2011]) to linear transformations in LMR. Then, we give characterizations of ss-goodness in the context of LMR. Using the two characteristic ss-goodness constants, γsγ_s and γ^s\hatγ_s, of a linear transformation, not only do we derive necessary and sufficient conditions for a linear transformation to be ss-good, but also provide sufficient conditions for exact and stable ss-rank matrix recovery via the nuclear norm minimization under mild assumptions. Moreover, we give computable upper bounds for one of the ss-goodness characteristics which leads to verifiable sufficient conditions for exact low-rank matrix recovery.

Open access
preprint2010arXiv

Homogeneous Cone Complementarity Problems and PP Properties

We consider existence and uniqueness properties of a solution to homogeneous cone complementarity problem (HCCP). Employing the TT-algebraic characterization of homogeneous cones, we generalize the P,P0,R0P, P_0, R_0 properties for a nonlinear function associated with the standard nonlinear complementarity problem to the setting of HCCP. We prove that if a continuous function has either the order-P0P_0 and R0R_0, or the P0P_0 and R0R_0 properties then all the associated HCCPs have solutions. In particular, if a continuous function has the trace-PP property then the associated HCCP has a unique solution (if any); if it has the uniform-trace-PP property then the associated HCCP has the global uniqueness (of the solution) property (GUS). We present a necessary condition for a nonlinear transformation to have the GUS property. Moreover, we establish a global error bound for the HCCP with the uniform-trace-PP property. Finally, we study the HCCP with the relaxation transformation on a TT-algebra and automorphism invariant properties for homogeneous cone linear complementarity problem.

Open access

Research graph

See the researcher in context

Open full explorer

Inspect adjacent papers, topics, institutions and collaborators without losing the researcher page.

Building this map preview

BZPEER is loading the nearby papers, people, topics and institutions for this page.