网络结构

Moreover the structure of the neural network is further optimized, so the convergence rate of the algorithm is obviously accelerated. At last the robusticity of the algorithm is carefully inspected and the result is perfect, and is worth of research.

在此基础上对网络结构进行了进一步的优化，提高了算法的收敛性，并且对算法的抗噪性进行了检验，结果表明算法的抗噪性良好，有较高的应用价值。

It can be seen that the size of the microgels is around 50～60 nm in diameter in the solid state,and the microgels exhibit a narrow size distribution with an average diameter of around 100 nm in the wet state.2.Pore size and morphology of hydrogelThe SEM images of freeze-dried hydrogels exhibited a highly macroporous spongelike structure and the average mesh size is about 10μm.

孔径通过扫描电子显微镜可以看到凝胶内部呈现多孔的类似海绵内部的结构，孔径在10μm左右，浸泡24h后可以发现网络结构发生了降解，孔径变大，达到20μm左右，浸泡7d后凝胶进一步降解，孔径增大到40～50μm.3。

Here, our focus is on the process of network structuration and its distributional implications, rather than on the individual egos that comprise the network.

我们这里关注的是网络结构化的过程及其分布的影响，而不是构成网络的个体自我。

The thesis based on Yellow River Diversion Project. Through the research of the existing design of the supervising system of Yellow River Diversion Project, this passage pointed out some problems hard to be dealt with as follows:(1) Owing to the dispersion of control points and remote transfers of data, it cost a lot to construct the network of the supervising system and it was hard to implement the construction.(2) The connection mode of network needed to be taken into acount. Once the configuration of network changed or the nodes increased, the supervising system needed to be reconfigured.(3) Owing to the tanglesome physical communication line, the man-made communication error happened frequently.(4) When the communication malfunction occured, it needed various engineers to check and repair the fault.(5) It cost a lot to operate and maintain the supervising system.

本文以世界银行援建项目&Yellow River Diversion Project&(引黄入晋工程，编号：0015-K81062)为背景，在对引黄入晋工程的远程监控系统进行研究分析的过程中，指出了该系统在目前采用的有线光纤和有线电缆通讯技术下难以处理的一些问题：(1)由于监控点分散且传输距离较远，有线光纤和有线电缆的网络建设成本很高，实施困难；(2)需要考虑节点之间的连接形式，一旦网络结构发生变化或节点需要扩充，监控系统便需要调整，耗费大量的人力物力；(3)由于系统庞大，通讯的物理线路较为繁乱，在建设和维护阶段容易出现人为触碰等引起的通讯故障；(4)在出现通讯故障时，需要多方人员配合进行故障检测，不仅耗费人力和物力，且不利于故障排除；(5)运行和维护成本较高，需要很多运行维护人员进行大量繁杂的工作。

Up to now, the whole operational speed of existent CNN chips has reached Tera magnitude, which can great improve the speed and real time of video processing.

细胞神经网络在视频运动对象分割中的应用研究研究细胞神经网络用于视频运动对象分割的原因在于其并行高速的特点和适于图象处理的网络结构。

In this paper, we propose a novel method of inferring node topologic ordering, which makes up for the deficiency of Bayesian networks needing known node ordering.

本文提出了一种推断节点拓扑顺序的方法，它弥补了贝叶斯网络结构学习需要节点顺序的不足。

All carboxylates from acetate ligands coordinated in tridentate mode and bridge two M centers in adjacent different sheets.

而通过乙酸根配体以三齿模式桥联相邻近的毯子状网络结构的两个金属离子而形成三维网状结构。

In the 1990s, a more complete theoretical system - the achievement of statistical learning theory and neural networks, and other than because of the emerging machine learning methods of the difficulties encountered some important, such as how to determine the issue of network structure, overfitting and underfitting, the local minimum points ect, makes the rapid development and improvement of SVM in resolving the small sample, non-linear and high-dimensional pattern recognition problems in the performance of many unique advantages and can function to promote the use of fitted function and other machines learning problems.

上世纪90年代，一个较完善的理论体系--统计学习理论( Statistical Learning Theory，简称SLT)的实现和由于神经网络等较新兴的机器学习方法的研究遇到一些重要的困难，比如如何确定网络结构的问题、过学习与欠学习问题、局部极小点问题等，使得SVM迅速发展和完善，在解决小样本、非线性及高维模式识别问题中表现出许多特有的优势，并能够推广应用到函数拟合等其他机器学习问题中。

The algorithm's main feature is to maintain the fixed length of data by using glide window technology and avoid the data being saturated. The Pruning technology was used to make the network architecture unsustained growth, and avoid the fitting and the training phenomenon, thus, it has guaranteed system's pan-ability.

其算法的最主要特点是：使用滑动窗技术保持固定长度的数据来调整参数，避免了数据饱和；使用了修剪技术，使得网络结构不会持续增长，避免了过拟合及过训练现象，因而确保了系统的泛化能力。

IEPA was combined with neural network by the new algorithm, multi-layer feed-forward neural network structure was replaced with RBF neural network structure, and back propagation algorithm was replaced with IEPA, then, expected side-lobe suppression quota was rapidly waisted.

该算法将进化规划算法和神经网络结合起来，用径向基RBF神经网络结构取代多层前馈网络MFNN（Multilayer Feedforward Neural Nerworks）结构，用改进进化规划取代反向传播算法BP，可以以较大概率快速的收敛到要求的旁瓣抑制指标。

However, as the name（read－only memory）implies, CD disks cannot be written onorchanged in any way.

然而，正如其名字所指出的那样，CD盘不能写，也不能用任何方式改变其内容。

Galvanizes steel pallet is mainly export which suits standard packing of European Union, the North America. galvanizes steel pallet is suitable to heavy rack. Pallet surface can design plate type, corrugated and the gap form, satisfies the different requirements.

镀锌钢托盘多用于出口，替代木托盘，免薰蒸，符合欧盟、北美各国对出口货物包装材料的法令要求；喷涂钢托盘适用于重载上货架之用，托盘表面根据需要制作成平板状、波纹状及间隔形式，满足不同的使用要求。