一种插值新方法

研究了数值计算方法教学中的插值问题.利用再生核希尔伯特空间中再生核的再生性,获得了一维插值问题的新方法,通过数值算例验证了所提出的插值方法的有效性.     

计算再生核的Green函数方法

本文讨论了利用Green函数计算再生核的方法,在Wm2空间中利用再生核的和性质以及Green函数理论给出再生核构造的一般方法,并利用此方法计算出W32空间的再生核.     

再生核空间中一类非线性积分方程的求解方法

本文在再生核空间中,利用再生核把非线性积分方程化为线性积分方程,研究了此类方程的求解问题,揭示了此类方程解的结构,存在性及多解等问题.     

求解奇异摄动抛物型对流扩散问题的再生核方法（英文）

基于域分解方法和再生核方法,文章提出了一种求解一维奇异摄动抛物型对流扩散问题的数值方法.原问题被分解成边界层区域问题和正则区域问题,正则区域问题的近似解通过原问题对应的退化问题的解进行近似,边界层区域问题的近似解通过构造合适的再生核,并利用再生核理论给出.三个数值算例的实验结果表明所提出的数值方法是有效的.     

高斯核正则化学习算法的泛化误差

对广义凸损失函数和变高斯核情形下正则化学习算法的泛化性能展开研究.其目标是给出学习算法泛化误差的一个较为满意上界.泛化误差可以利用正则误差和样本误差来测定.基于高斯核的特性,通过构构建一个径向基函数(简记为RBF)神经网络,给出了正则误差的上界估计,通过投影算子和再生高斯核希尔伯特空间的覆盖数给出样本误差的上界估计.所获结果表明,通过适当选取参数σ和λ,可以提高学习算法的泛化性能.     

关于W_2~1[a,b]能否扩大的一些问题

讨论了W12[a,b]能否扩大为含有有间断点函数的再生核空间的问题.结论是:若再生核空间W W12[a,b]含有有间断点的函数,则间断点必固定、间断点个数必有限且非端点a,b.进一步,我们构造了函数含有n个间断点的再生核空间并给出其再生核表达式.     

Cgau小波变换像空间中的等距变换与反演公式

在以再生核Hilbert空间为连续小波变换像空间的基础上,针对Cgau小波(复数形式的Gauss小波),给出了其小波变换像空间的再生核的具体表达式.当固定尺度因子时,利用再生核空间理论,对Cgau小波变换像空间做了具体描述,分别给出了Cgau小波变换像空间中的等距变换和反演公式,这为进一步研究一般的小波变换像空间提供了理论基础.     

Cgau小波变换像空间中的等距变换与反演公式

在以再生核Hilbert空间为连续小波变换像空间的基础上,针对Cgau小波(复数形式的Gauss小波),给出了其小波变换像空间的再生核的具体表达式.当固定尺度因子时,利用再生核空间理论,对Cgau小波变换像空间做了具体描述,分别给出了Cgau小波变换像空间中的等距变换和反演公式,这为进一步研究一般的小波变换像空间提供了理论基础.     

W_2~m[a,b]空间中线性泛函最佳逼近的深入讨论

本文基础上给出了Wm2(a,b)空间再生核构造的普遍方法,并利用再生核讨论算于插值样条的投影性质及最佳数值逼近问题     

Journe小波变换像空间描述

1引言1950年N.Aronszjan发表的一篇综述性文章《Theory of reproducing kernels》标志再生核理论的初步形成.由于再生核有许多良好的计算性质,S.Saitoh总结并深入研究再生核基本理论,进一步拓展了再生核的应用领域;徐利治把再生核应用于L~2(B)(B是复平面上的一个区域)中解析函数重积分降维问题,并提出了一个能对一些预先给出的     

Application of reproducing kernel method to third order three-point boundary value problems

This paper investigates the analytical approximate solutions of third order three-point boundary value problems using reproducing kernel method. The solution obtained by using the method takes the form of a convergent series with easily computable components. However, the reproducing kernel method can not be used directly to solve third order three-point boundary value problems, since there is no method of obtaining reproducing kernel satisfying three-point boundary conditions. This paper presents a method for solving reproducing kernel satisfying three-point boundary conditions so that reproducing kernel method can be used to solve third order three-point boundary value problems. Results of numerical examples demonstrate that the method is quite accurate and efficient for singular second order three-point boundary value problems.     

Numerical algorithm for parabolic problems with non-classical conditions

The parabolic problems with non-classical conditions are discussed in a reproducing kernel space in this paper. A reproducing kernel space is constructed, in which the non-classical conditions of the parabolic problems are satisfied. Based on the reproducing kernel space, a new technique for solving the non-classical parabolic problems is presented. Some examples are displayed to demonstrate the validity and applicability of the proposed method.     

Average sampling and reconstruction for reproducing kernel stochastic signals

This paper mainly considers the problem of reconstructing a reproducing kernel stochastic signal from its average samples. First, a uniform convergence result for reconstructing the deterministic reproducing kernel signals by an iterative algorithm is established. Then, we prove that the quadratic sum of the corresponding reconstructed functions is uniformly bounded. Moreover, the reconstructed functions provide a frame expansion in the special case p = 2. Finally, the mean square convergence for recovering a weighted reproducing kernel stochastic signal from its average samples is given under some decay condition for the autocorrelation function, which can be removed for the case p = 2. Copyright © 2015 John Wiley & Sons, Ltd.     

A comparative analysis of optimization and generalization properties of two-layer neural network and random feature models under gradient descent dynamics

A fairly comprehensive analysis is presented for the gradient descent dynamics for training two-layer neural network models in the situation when the parameters in both layers are updated. General initialization schemes as well as general regimes for the network width and training data size are considered. In the overparametrized regime, it is shown that gradient descent dynamics can achieve zero training loss exponentially fast regardless of the quality of the labels. In addition, it is proved that throughout the training process the functions represented by the neural network model are uniformly close to those of a kernel method. For general values of the network width and training data size, sharp estimates of the generalization error are established for target functions in the appropriate reproducing kernel Hilbert space.     

Construction of cubature formulae with preassigned nodes§

In this article, a technique for developing cubature rules with preassigned nodes is presented to avoid wasting of information in scientific computation. The corresponding constructive method of the cubature rule is also given. As an application of the rules, a cubature formula on disk, which was derived via the method of reproducing kernel in (Xu, Y., 2000, Constructing cubature formulae by the method of reproducing kernel. Numerische Mathematik, 85, 155–173), is reconstructed by using our technique. When the preassigned nodes are selected as the nodes of a cubature formula of lower degree, an embedded cubature formula can be easily obtained by the presented method. Furthermore, some examples are included in the article.     

再生核空间W_2~1[a,b]中线性微分方程组初值问题的数值求解

再生核方法在某些方程(组)解的表示和逼近中具有独特的优势,用一种新的再生核讨论线性微分方程组初值问题解的精确表示与近似计算.较之以往的同类文章,对一些重要定理进行了更简单有效的证明.另外本文的再生核由于其结构简单,易于算法实现.最后的算例充分的显示了基于的再生核方法的有效性.     

Simplified reproducing kernel method for fractional differential equations with delay

This paper is devoted to the numerical scheme for the delay initial value problems of a fractional order. The main idea of this method is to establish a novel reproducing kernel space that satisfies the initial conditions. Based on the properties of the new reproducing kernel space, the simplified reproducing kernel method (SRKM for short) is applied to obtain accurate approximation. The Schmidt orthogoralization process which requires a large number of calculation is less likely to be employed. Numerical experiments are provided to illustrate the performance of the method.     

Sampling in reproducing kernel Banach spaces on Lie groups

We present sampling theorems for reproducing kernel Banach spaces on Lie groups. Recent approaches to this problem rely on integrability of the kernel and its local oscillations. In this paper, we replace these integrability conditions by requirements on the derivatives of the reproducing kernel and, in particular, oscillation estimates are found using derivatives of the reproducing kernel. This provides a convenient path to sampling results on reproducing kernel Banach spaces. Finally, these results are used to obtain frames and atomic decompositions for Banach spaces of distributions stemming from a cyclic representation. It is shown that this process is particularly easy, when the cyclic vector is a Gårding vector for a square integrable representation.     

Numerical solution of nonlinear three‐point boundary value problem on the positive half‐line

In this paper, we present an efficient numerical algorithm to solve the three‐point boundary value problem on the half‐line based on the reproducing kernel theorem. Considering the boundary conditions including a limit form, a new weighted reproducing kernel space is established to overcome the difficulty. By applying reproducing property and existence of the orthogonal basis in the weighted reproducing kernel space, the approximate solution is constructed by the orthogonal projection of the exact solution. Convergence has also been discussed. We demonstrate the accuracy of the method by numerical experiments. Copyright © 2012 John Wiley & Sons, Ltd.