自然梯度算法是其中非常重要的算法之一。
Natural gradient algorithm is one of the most important algorithms.
本文基于互信息极小提出了ICA的梯度算法。
In this paper, a simple gradient algorithm of ICA is developed using minimum mutual information.
然而，变步长自然梯度算法性能仍然不够理想。
However, the performance of variable step-size natural gradient algorithm is still less than ideal.
同时，还可使用迭代的预条件双共轭梯度算法进行计算。
An iterative conjugate gradient algorithm is used in all cases.
丈中利用双梯度算法对自然图像的基向量进行迭代学习。
The basis vectors of the natural images were obtained by using fast conjugate gradient algorithm.
此外，相应于次梯度算法的实现方式，提出两种分布协调结构。
In addition, two coordination frameworks were proposed corresponding to the two implementing ways of sub-gradient algorithm.
本文证明了使用梯度算法，新价值函数的全局收敛性能得到保证。
By using gradient algorithm, the global convergence of the cost-function can be guaranteed.
分离系统的线性部分和非线性部分参数学习都采用自然梯度算法。
The natural gradient method is applied for parameter learning of the linear and nonlinear parts of the separating system.
对于单输出前馈神经网络的梯度算法的收敛性已经有了详细的讨论。
The convergence of the gradient algorithm for feedforward neural networks with one output unit has thoroughly studied.
研制的最小平方共轭梯度算法和奇异值分解法也可进行同样的层析反演。
The least square conjugate gradient and odd values decomposition method also can be used to perform tomographic inversion.
提出一种稳定的IR自适应滤波器，并给出了系数迭代的非线性梯度算法。
This paper puts forwards a stable IIR adaptive filter, along with the nonlinear gradient algorithm concerning its coefficient iteration.
提出一种交替梯度算法，对径向基函数(RBF)神经网络的训练进行改进。
One kind of alternant gradient algorithm for improving the training of Radial Basis Function (RBF) neural network is proposed.
仿真结果显示，自然梯度算法比传统梯度算法收敛速度更快，分离效果更好。
Simulation results show that the natural gradient approach has faster convergence speed and better separation performance than the conventional gradient based algorithm.
本算法用自适应算法，二次型共轭梯度算法，解前后向预测构成的矩阵方程。
The new algorithm uses the quadratic conjugate gradient algorithm (QCGA) to solve the forward-backward linear prediction matrix equation.
提出了一种基于几何原理求解结构可靠指标的高效实用算法，即随机梯度算法。
An efficient and practical stochastic gradient algorithm for calculating structural reliability index is presented in this paper on basis of geometrical principle of reliability index.
根据无约束最优化问题的梯度算法，提出了二次梯度算法，并证明了其收敛性。
In this paper, the gradient computational algorithm about unconditional extreme value is given.
本文在介绍了盲源分离的基础理论的基础上，对自然梯度算法进行了详细的介绍。
This paper describes the basic theory of blind source separation and natural gradient algorithm in detail.
当相位梯度算法挑选的散射点不是孤立散射点时，导致相位误差估计的准确性下降。
If the scatter selected in phase gradient algorithm(PGA) algorithm is not isolated, the accuracy of the phase error estimation will be degraded.
本文使用信赖域策略结合投影梯度算法来解约束优化问题，并给出算法及其收敛性。
This paper is to study the convergence properties of the gradient projection method with trust region strategy for constrained optimization.
提出了径流长期分级预报的人工神经网络方法，给出了多层前馈网络的联合梯度算法。
An artificial neural network method is proposed for hydrological forecasting, the hybrid gradient method for multilayer feedforward neural network developed.
自然梯度算法有收敛速度和稳定误差两个重要指标，然而这两个指标存在内在的矛盾。
In the natural gradient algorithm, the convergence speed and stability error are two important indicators, but these two indicators are of inherent contradictions.
本文以递推随机梯度算法为基础，给出了寻求一类非线性系统的自适应控制律的方法。
In this paper, on the base of the recursive stochastic gradient method, a kind of adaptive control algorithm of the non-linear system has been given.
在常规算法的基础上，提出了一种基于变斜率算法与共轭梯度算法的接力逆向传播算法。
An artificial neural network algorithm, based on variable slope BP algorithm and conjugate gradient algorithm, is presented in this paper.
该算法采用了相位梯度算法(PGA)的处理结构，利用了多个距离单元上的散射点信号。
In the algorithm, the phase gradient algorithm (PGA) was performed, and the signals of multiple scatters on a few range bins were utilized.
根据纹理位于图像中灰度突变处的特点，采用典型梯度算法，对目标图像进行两次梯度锐化。
According to the feature that the surface textures locate in the sudden change areas of image grey value, the double grade sharpening was performed for the image using the typical grad algorithm.
针对闪光照相系统成像信噪比低的特点，提出了正则化预优约束共轭梯度算法(RPCCG)。
Based on the constrained conjugate gradient algorithm, a regularization preconditioned constrained conjugate gradient algorithm (RPCCG) was proposed for image-reconstruction for radiography.
基于梯度算法和前馈网络所具有的普遍近似性质，提出了一种新的监督型多目标系统化训练机制。
Based on gradient algorithm and the fundamental approximation of feedforward network, a new supervised comprehensive training mechanism is put forward.
将梯度优化过程看作反馈控制系统，从而提出了一种带动量项的PID梯度算法(PIDGDM)。
An algorithm of PID gradient descent with momentum term (PIDGDM) is proposed. In this algorithm, the procedure of gradient optimization is considered as a feedback control system.
本文对该算法中的线搜索进行了推广，提出了一种新的非线性共轭梯度算法并证明了其全局收敛性。
This paper presents a wide line search and gives a new conjugate gradient method, with global convergence.
本文对该算法中的线搜索进行了推广，提出了一种新的非线性共轭梯度算法并证明了其全局收敛性。
This paper presents a wide line search and gives a new conjugate gradient method, with global convergence.
应用推荐