同伦摄动稀疏正则化方法及其应用

稀疏正则化方法在参数重构中起到了越来越重要的作用.与传统的正则化方法相比,稀疏正则化方法能较好地重构稀疏变量.由于稀疏正则化的不可微性,需要对已有的经典算法进行改进.本文构建同伦摄动稀疏正则化方法克服标准稀疏正则化的不可微性,并将该方法应用到基于布莱克一斯科尔斯期权定价模型重构隐含波动率和基于托达罗模型重构政策参数.数值实验表明,所提出的方法是收敛和稳定的.     

一种基于正则化的稀疏表示方法

提出了一种基于正则化技术的信号稀疏表示方法.该方法与经典稀疏表示算法的主要区别可概括为两点:其一,直接使用e_0模而不是被广泛采用的e_1模来度量稀疏性;其二,正则化项的引入使得该模型得到的信号表达是所有表示中最优稀疏的.在本文中,正则化项采用框架势来描述稀疏表示的"最优性",利用二次可微的凹函数来逼近e_0模,得到了求解所提出的正则化模型的近似算法,并给出了收敛性分析.此外,数值实验也显现了本文所提模型及算法相比于经典算法的优越性.     

稀疏优化模型及其正则化方法

稀疏优化模型是目前最优化领域中非常热门的研究前沿课题,在压缩感知、图像处理、机器学习和统计建模等领域都获得了成功的应用.本文以光谱分析技术、数字信号处理和推荐系统等多个应用问题为例,阐述稀疏优化模型的建模过程与核心思想.稀疏优化模型属于组合优化模型,非常难以求解(NP-难).正则化方法是稀疏优化模型的一类常用的求解方法.我们将介绍正则化方法的原理与几类常见的正则化模型,并阐述正则化模型的稳定性理论与多种先进算法.数值实验表明,这些算法都具有快速、高效、稳健等显著优点.稀疏正则化模型将在大数据时代中发挥更显著的计算优势与应用价值.     

求解波场变换反演问题的连续正则化方法

本文研究波场变换反演问题.利用连续正则化方法求解波场变换反演问题,构造展平泛函,基于已经正则化的变分问题用差分法作有限维逼近.利用偏差原理和Newton三阶迭代收敛格式选出最优的正则化参数,实施数值求解.通过对数值计算结果与已知波场函数对比,证明该方法的有效性和可行性.与离散正则化算法相比,本文的连续正则化算法具有保结构和收敛速度快等优点.     

一个求解消失约束数学规划的部分光滑正则化方法（英文）

本文研究带有消失约束的数学规划问题.针对这一问题,我们提出了一种基于伪Huber函数的光滑正则化方法,该方法只对部分消失约束进行光滑化.对于新的光滑问题,我们证明Mangasarian-Fromovitz约束规格在某些情况下是成立的.我们也分析该方法的收敛性质,即,一个光滑正则化问题稳定点序列的聚点是原问题的T-稳定点,并给出光滑正则化问题稳定点序列的聚点是原问题的M-稳定点或S-稳定点的一些充分条件.最后初步的数值结果表明该方法是可行的.     

电离层层析成像及数值模拟

利用正则化反演法,对电离层层析成像技术进行了数值模拟.模拟结果表明,正则化反演法能够准确反演出电离层赤道异常的电子密度分布.对正则化反演误差的原因进行了分析,定义了代表重建结果精度的参数并研究其随信号数量的变化,结果显示造成反演误差的主要原因是信息量不足,在提供充足的数据条件下,正则化反演法能够得到很高精度的重建结果.数值模拟实验进一步证明了电离层层析技术在反演电离层电子密度分布中有很大的应用价值.     

关于一类广义Tikhonov 正则化方法的饱和效应分析

Tikhonov正则化方法是研究不适定问题最重要的正则化方法之一,但由于这种方法的饱和效应出现的太早,使得无法随着对解的光滑性假设的提高而提高正则逼近解的收敛率,也即对高的光滑性假设,正则解与准确解的误差估计不可能达到阶数最优.Schrroter T 和Tautenhahn U给出了一类广义Tikhonov正则化方法并重点讨论了它的最优误差估计, 但却未能对该方法的饱和效应进行研究.本文对此进行了仔细分析,并发现此方法可以防止饱和效应,而且数值试验结果表明此方法计算效果良好.     

用于神经网络权值稀疏化的L_(1/2)正则化方法

在保证适当学习精度前提下,神经网络的神经元个数应该尽可能少(结构稀疏化),从而降低成本,提高稳健性和推广精度.本文采用正则化方法研究前馈神经网络的结构稀疏化.除了传统的用于稀疏化的L1正则化之外,本文主要采用近几年流行的L1/2正则化.为了解决L1/2正则化算子不光滑、容易导致迭代过程振荡这一问题,本文试图在不光滑点的一个小邻域内采用磨光技巧,构造一种光滑化L1/2正则化算子,希望达到比L1正则化更高的稀疏化效率.本文综述了近年来作者在用于神经网络稀疏化的L1/2正则化的一些工作,涉及的神经网络包括BP前馈神经网络、高阶神经网络、双并行前馈神经网络,以及Takagi-Sugeno模糊模型.     

稀疏观测数据下热传导方程逆时反问题的数值反演

探讨一类带第二类边界条件的一维热传导方程逆时问题,首先利用分离变量法推导了反问题的积分表达形式,然后基于解析延拓技术,证明了基于稀疏附加数据下反问题的唯一性,并对反问题的不适定性进行说明,接着利用线性叠加原理及有限元插值技术,给出了该逆时反演问题对应的离散反演方程组形式,借助于Tikhonov正则化方法和正则化参数选取的广义交叉验证准则,设计出了该逆时反演问题的直接反演算法,最后通过数值算例说明所设计的直接反演算法是有效的.     

基于卫星云图的风矢场度量模型与算法探讨

首先通过换算视场坐标确定灰度矩阵中每个元素对应的采样点在地球上的经纬度,从而将灰度矩阵转化为卫星云图,并添加海岸线.在此基础上,使用相关匹配法对具有一定时间间隔的两幅相关卫星云图进行模板匹配生成风矢场(云导风)矢量.然后,借助于近年来发展起来的数值微分方法,从图像灰度中提取出图像梯度信息,再利用正则化方法,实现了云导风的反演.对云图中加入灰度梯度信息和未加入灰度梯度信息的风场反演结果进行比较.结果表明,加入图像灰度梯度信息后所实施的新反演方法可有效减小图像干扰的影响,同时也大大提高了风矢量反演的精度,为卫星云图反演云导风探索出一条新路.     

Sparsity reconstruction in electrical impedance tomography: An experimental evaluation

We investigate the potential of sparsity constraints in the electrical impedance tomography (EIT) inverse problem of inferring the distributed conductivity based on boundary potential measurements. In sparsity reconstruction, inhomogeneities of the conductivity are a priori assumed to be sparse with respect to a certain basis. This prior information is incorporated into a Tikhonov-type functional by including a sparsity-promoting ?¹-penalty term. The functional is minimized with an iterative soft shrinkage-type algorithm. In this paper, the feasibility of the sparsity reconstruction approach is evaluated by experimental data from water tank measurements. The reconstructions are computed both with sparsity constraints and with a more conventional smoothness regularization approach. The results verify that the adoption of ?¹-type constraints can enhance the quality of EIT reconstructions: in most of the test cases the reconstructions with sparsity constraints are both qualitatively and quantitatively more feasible than that with the smoothness constraint.     

A Joint Tikhonov Regularization and Augmented Lagrange Approach for Ill-Posed State Constrained Control Problems with Sparse Controls

Abstract

We provide a modified augmented Lagrange method coupled with a Tikhonov regularization for solving ill-posed state constrained elliptic optimal control problems with sparse controls. We consider a linear quadratic optimal control problem without any additional L² regularization terms. The sparsity is guaranteed by an additional L¹ term. Here, the modification of the classical augmented Lagrange method guarantees us uniform boundedness of the multiplier that corresponds to the state constraints. We present a coupling between the regularization parameter introduced by the Tikhonov regularization and the penalty parameter from the augmented Lagrange method, which allows us to prove strong convergence of the controls and their corresponding states. Moreover, convergence results proving the weak convergence of the adjoint state and weak*-convergence of the multiplier are provided. Finally, we demonstrate our method in several numerical examples.     

组模偏正则化及其应用

本文研究组模下偏正则最小化问题,证明了解的存在性,稀疏性.研究了零空间性质对最优解的刻画.仔细探讨了解的一种单调性,并应用这种单调性说明最优化问题的求解可以分解到各组中.最后给出了一个所证定理在地震反演的应用.     

Video deraining via nonlocal low-rank regularization

Outdoor videos captured in rainy weather may be significantly corrupted by the undesired rain streaks, which severely affect the video processing tasks in outdoor computer vision systems. In this paper, we propose a tensor-based video rain streaks removal method using the nonlocal low-rank regularization. Specifically, we first divide videos into overlapped spatial–temporal patches. Then for each patch, we group its nonlocal similar spatial–temporal patches to form a third-order tensor. To model the clean videos, we characterize the wealth redundancy by adopting the tensor nuclear norm to regularize the low-rankness of the third-order tensors formed by similar spatial–temporal patches of clean videos. We also consider the piecewise smoothness and the temporal continuity of clean videos and utilize the unidirectional total variation to enhance the smoothness and continuity. Moreover, as rain streaks are sparse and smooth along the rain direction, we model the rain streaks by employing an ℓ₁ norm and the unidirectional total variation penalty to boost the sparsity and directional smoothness, respectively. We develop an efficient alternating direction method of multipliers to solve the proposed model. Experimental results on both synthetic and real rainy videos show that our method outperforms the state-of-the-art methods quantitatively and qualitatively.     

ACCURATE AND EFFICIENT IMAGE RECONSTRUCTION FROM MULTIPLE MEASUREMENTS OF FOURIER SAMPLES

Several problems in imaging acquire multiple measurement vectors (MMVs) of Fourier samples for the same underlying scene. Image recovery techniques from MMVs aim to exploit the joint sparsity across the measurements in the sparse domain. This is typically accomplished by extending the use of $\ell_1$ regularization of the sparse domain in the single measurement vector (SMV) case to using $\ell_{2,1}$ regularization so that the "jointness" can be accounted for. Although effective, the approach is inherently coupled and therefore computationally inefficient. The method also does not consider current approaches in the SMV case that use spatially varying weighted $\ell_1$ regularization term. The recently introduced variance based joint sparsity (VBJS) recovery method uses the variance across the measurements in the sparse domain to produce a weighted MMV method that is more accurate and more efficient than the standard $\ell_{2,1}$ approach. The efficiency is due to the decoupling of the measurement vectors, with the increased accuracy resulting from the spatially varying weight. Motivated by these results, this paper introduces a new technique to even further reduce computational cost by eliminating the requirement to first approximate the underlying image in order to construct the weights. Eliminating this preprocessing step moreover reduces the amount of information lost from the data, so that our method is more accurate. Numerical examples provided in the paper verify these benefits.     

Regularized collocation method for Fredholm integral equations of the first kind

In this paper we consider a collocation method for solving Fredholm integral equations of the first kind, which is known to be an ill-posed problem. An “unregularized” use of this method can give reliable results in the case when the rate at which smallest singular values of the collocation matrices decrease is known a priori. In this case the number of collocation points plays the role of a regularization parameter. If the a priori information mentioned above is not available, then a combination of collocation with Tikhonov regularization can be the method of choice. We analyze such regularized collocation in a rather general setting, when a solution smoothness is given as a source condition with an operator monotone index function. This setting covers all types of smoothness studied so far in the theory of Tikhonov regularization. One more issue discussed in this paper is an a posteriori choice of the regularization parameter, which allows us to reach an optimal order of accuracy for deterministic noise model without any knowledge of solution smoothness.     

Adaptive iterative thresholding algorithms for magnetoencephalography (MEG)

We provide fast and accurate adaptive algorithms for the spatial resolution of current densities in MEG. We assume that vector components of the current densities possess a sparse expansion with respect to preassigned wavelets. Additionally, different components may also exhibit common sparsity patterns. We model MEG as an inverse problem with joint sparsity constraints, promoting the coupling of non-vanishing components. We show how to compute solutions of the MEG linear inverse problem by iterative thresholded Landweber schemes. The resulting adaptive scheme is fast, robust, and significantly outperforms the classical Tikhonov regularization in resolving sparse current densities. Numerical examples are included.     

Minimal Zero Norm Solutions of Linear Complementarity Problems

In this paper, we study minimal zero norm solutions of the linear complementarity problems, defined as the solutions with smallest cardinality. Minimal zero norm solutions are often desired in some real applications such as bimatrix game and portfolio selection. We first show the uniqueness of the minimal zero norm solution for Z-matrix linear complementarity problems. To find minimal zero norm solutions is equivalent to solve a difficult zero norm minimization problem with linear complementarity constraints. We then propose a p norm regularized minimization model with p in the open interval from zero to one, and show that it can approximate minimal zero norm solutions very well by sequentially decreasing the regularization parameter. We establish a threshold lower bound for any nonzero entry in its local minimizers, that can be used to identify zero entries precisely in computed solutions. We also consider the choice of regularization parameter to get desired sparsity. Based on the theoretical results, we design a sequential smoothing gradient method to solve the model. Numerical results demonstrate that the sequential smoothing gradient method can effectively solve the regularized model and get minimal zero norm solutions of linear complementarity problems.     

Numerical solution of forward and backward problem for 2-D heat conduction equation

For a two-dimensional heat conduction problem, we consider its initial boundary value problem and the related inverse problem of determining the initial temperature distribution from transient temperature measurements. The conditional stability for this inverse problem and the error analysis for the Tikhonov regularization are presented. An implicit inversion method, which is based on the regularization technique and the successive over-relaxation (SOR) iteration process, is established. Due to the explicit difference scheme for a direct heat problem developed in this paper, the inversion process is very efficient, while the application of SOR technique makes our inversion convergent rapidly. Numerical results illustrating our method are also given.