一种偏好于查询子流形的半监督图像检索算法

在基于反馈的图像检索中,由于被用户标记为相关和不相关的图像数较少,使得检索问题变成了一个典型的小样本问题.流形可表达数据在低维空间中的内在几何结构,流形正则化的目的是利用这种几何结构来约束解空间,以使最优解能反映数据本身的几何分布.为了解决反馈检索中的小样本问题,本文在流形正则化框架下提出一个新的半监督图像检索算法.在新算法中,流形正则化项只依赖于文中定义的查询子流形,而不依赖于数据集的全局结构.在两个图像集上的实验结果对比表明,本文提出的新算法在检索效果上优于现有的4种state-of-the-art算法.     

基于混合最小二乘支持向量机网络模型的非线性系统辨识

针对基于输入输出数据的非线性系统辨识问题,提出一种新的混合最小二乘支持向量机(LS-SVMs)网络模型及相应的学习算法.该算法将系统的辨识问题动态自适应的划分为若干子问题,将支持向量机(SVM)用于各子模块辨识;通过分析模型的统计学特性,给出基于整体框架优化的系统参数辨识方法.针对系统中参数相关联的特性,采用期望条件最大化(ECM)算法对其进行条件辨识,同时结合正则化理论和最小二乘法,保证各专家模块的结构风险最小化辨识原则.试验结果表明,该方法兼具良好的辨识精度和泛化性能.     

Numerical inversions of a source term in the FADE with a Dirichlet boundary condition using final observations

This paper deals with an inverse problem of determining a source term in the one-dimensional fractional advection-dispersion equation (FADE) with a Dirichlet boundary condition on a finite domain, using final observations. On the basis of the shifted Grünwald formula, a finite difference scheme for the forward problem of the FADE is given, by means of which the source magnitude depending upon the space variable is reconstructed numerically by applying an optimal perturbation regularization algorithm. Numerical inversions with noisy data are carried out for the unknowns taking three functional forms: polynomials, trigonometric functions and index functions. The reconstruction results show that the inversion algorithm is efficient for the inverse problem of determining source terms in a FADE, and the algorithm is also stable for additional data having random noises.     

改进的二值图像复原方法研究

为了解决传统基于GBB方法的正则化二值图像复原算法放大背景噪声的问题,提出了一种基于GBB方法的改进的正则化二值图像复原算法。根据提出的算法进行实验,定性的对两种算法的图像复原结果进行了比较。同时,利用ISNR选取最佳参数,且在不同程度噪声的影响下定量的评价了两种复原方法的结果。实验结果表明,该方法对于模糊二值图像的复原效果有显著提升,有效降低了噪声对图像复原的影响。     

The geometric element transformation method for mixed mesh smoothing

The geometric element transformation method (GETMe) is a geometry-based smoothing method for mixed and non-mixed meshes. It is based on a simple geometric transformation applicable to elements bounded by polygons with an arbitrary number of nodes. The transformation, if applied iteratively, leads to a regularization of the polygons. Global mesh smoothing is accomplished by averaging the new node positions obtained by local element transformations. Thereby, the choice of transformation parameters as well as averaging weights can be based on the element quality which leads to high quality results. In this paper, a concept of an enhanced transformation approach is presented and a proof for the regularizing effect of the transformation based on eigenpolygons is given. Numerical examples confirm that the GETMe approach leads to superior mesh quality if compared to other geometry-based methods. In terms of quality it can even compete with optimization-based techniques, despite being conceptually significantly simpler.     

Experience-consistent modeling: Regression and classification problems

In this study, we are concerned with system modeling which involves limited data and reconciles the developed model with some previously acquired domain knowledge being captured in the format of already constructed models. Each of these previously available models was formed on a basis of extensive data sets which are not available for the current identification pursuits. To emphasize the nature of modeling being guided by the reconciliation mechanisms, we refer to this mode of identification as experience-consistent modeling. The paper presents the conceptual and algorithmic framework by focusing on regression models. By forming a certain extended form of the performance index, it is shown that the domain knowledge captured by regression models can play a similar role as a regularization component used quite commonly in system identification. Experimental results involve both synthetic low-dimensional data and selected data coming from Machine Learning repository. The data used in the experiments tackle regression models as well as classification problems (two-class classifiers).     

Some First-Order Algorithms for Total Variation Based Image Restoration

This paper deals with first-order numerical schemes for image restoration. These schemes rely on a duality-based algorithm proposed in 1979 by Bermùdez and Moreno. This is an old and forgotten algorithm that is revealed wider than recent schemes (such as the Chambolle projection algorithm) and able to improve contemporary schemes. Total variation regularization and smoothed total variation regularization are investigated. Algorithms are presented for such regularizations in image restoration. We prove the convergence of all the proposed schemes. We illustrate our study with numerous numerical examples. We make some comparisons with a class of efficient algorithms (proved to be optimal among first-order numerical schemes) recently introduced by Y. Nesterov.     

新型光滑正则半监督SVM方法及其在信用评级中的应用

提出了一种基于光滑正则的半监督支持向量机方法,并将其用于建立中小信用评级模型.它从少量标签样本和大量无标签样本中构造反映数据流形结构的光滑正则项,并结合到支持向量机的最大间隔分类器的学习过程.然后,提出一种渐进式方法来迭代获得“半标签”样本,稳健地提升支持向量机的泛化性能.在真实数据集上的实验结果表明,新方法获得的测试精度显著优于多种现有方法,非常适用于中小企业的信用评级任务.     

弹性多核学习

多核学习 (MKL) 的提出是为了解决多个核矩阵的融合问题, 多核学习求解关于多个核矩阵的最优的线性组合并同时解出对应于这个组合矩阵的支持向量机(SVM)问题. 现有的多核学习的框架倾向于寻找稀疏的组合系数, 但是当有信息的核的比例较高的时候, 对稀疏性的倾向会使得只有少量的核被选中而损失相当的分类信息. 在本文中, 我们提出了弹性多核学习的框架来实现自适应的多核学习. 弹性多核学习的框架利用了一个混合正则化函数来均衡稀疏性和非稀疏性, 多核学习和支持向量机问题都可以视作弹性多核学习的特殊情形. 基于针对多核学习的梯度下降法, 我们导出了针对弹性多核学习的梯度下降法. 仿真数据的结果显示了弹性多核学习方法相对多核学习和支持向量机的优势; 我们还进一步将弹性多核学习应用于基因集合分析问题并取得了有意义的结果; 最后, 我们比较研究了弹性多核学习与另一种利用了非稀疏思想的多核学习.