Instruction-level parallel processing

The performance of microprocessors has increased steadily over the past 20 years at a rate of about 50% per year. This is the cumulative result of architectural improvements as well as increases in circuit speed. Moreover, this improvement has been obtained in a transparent fashion, that is, without requiring programmers to rethink their algorithms and programs, thereby enabling the tremendous proliferation of computers that we see today. To continue this performance growth, microprocessor designers have incorporated instruction-level parallelism (ILP) into new designs. ILP utilizes the parallel execution ofthe lowest level computer operations-adds, multiplies, loads, and so on-to increase performance transparently. The use of ILP promises to make possible, within the next few years, microprocessors whose performance is many times that of a CRAY-IS. This article provides an overview of ILP, with an emphasis on ILP architectures-superscalar, VLIW, and dataflow processors-and the compiler techniques necessary to make ILP work well.     

神经网络及BP网络工作原理

1　神经网络概述神经网络是人们在模仿人脑处理问题的过程中发展起来的一种新型智能信息处理理论 ,它通过大量的称为神经元的简单处理单元构成非线性动力学系统 ,对人脑的形象思维、联想记忆等进行模拟和抽象 ,实现与人脑相似的学习、识别、记忆等信息处理能力。神经网络在经历了 4 0多年的曲折发展之后 ,在信息科学领域等许多应用方面已显示出巨大潜力和广阔的应用前景。由于神经网络的高速并行处理、分布存贮信息等特性符合人类视觉系统的基本工作原则 ,具有很强的自学习性、自组织性、容错性、高度非线性、高的鲁棒性、联想记忆功能和推…     

Neural processing of backwards-speech sounds

When sounds such as those produced by reverse playback of recorded speech are presented to left and right ears of normal subjects, the sounds arriving at the right ear are more accurately identified than those arriving at the left. These findings are comparable to the right-ear superiority demonstrated with normal speech sounds, and contrast with the left-ear superiority for musical and other non-speech sounds. It is suggested that the neural mechanisms underlying the perception of speech and non-speech sounds are not differentiated along the dimension of conceptual content.     

基于并联卷积神经网络的水果品种识别

为了解决传统的水果图像识别算法在特征提取上的缺陷,以及传统卷积神经网络识别率低的问题,设计了一种基于并联卷积神经网络来提取水果特征的识别方法,利用ELU激活函数替代ReLU激活函数,利用最大类间距损失函数结合传统SoftmaxWithLoss损失函数来提高对相似品种的识别准确率。选取Fruit-360数据集中的8个品种,利用边界均衡生成对抗网络(BEGAN)结合传统的数据增强方法生成大量高质量的数据集,并用其进行训练。结果表明,该模型对8个品种的平均识别准确率达98.85%,具有良好的识别效果。     

Neural network modelling of flat-plate solar collectors

In this study, different approaches to the modelling of flat-plate solar collectors are introduced and analysed. Among the physically based models, the heat network model and Hottel–Vhillier (H–V) models are discussed. The parameters of the latter model are identified for three different types of these solar collectors. The identification exhibited good agreement with the measured values. Finally, modelling simulations with an artificial neural network (ANN) technique were carried out. A sensitivity study was performed on the parameters of the neural network. The possible ANN structures, the size of training data set, the number of hidden neurons, and the type of training algorithm were analysed in order to identify the most appropriate model. The same ANN structures were trained and validated for the three solar collectors, using data generated from the H–V model and long-term (17 days) measurements.     

分布式数字资源中心网络存储系统的研究

如何将信息在网络中进行有效的存储和访问,是合理构建分布式数字资源中心的重要环节,本文对构建网络存储系统的主流技术进行较为系统的论述,并就构建分布式数字资源中心网络存储系统的几种解决方案进行了评析。     

微机网络并行计算环境Linux、并行计算平台MPI及其应用

介绍了并行计算研究的基本概况、微机网络并行计算环境－Ｌｉｎｕｘ操作系统及网络并行计算平台ＭＰＩ，并用实例说明微机网络并行计算环境和并行计算平台的建立和应用．     

马尾松人工林直径分布神经网络模型研究

该文首先以相对直径为输入变量,以累积频率为输出变量,构建了1∶S∶1的林分直径分布BP神经网络模型. 用1块具有代表性、26年生、全林伐倒测定每木胸径、树高生长过程的马尾松人工林标准地直径分布数据,作为马尾松人工林的期望分布,对所建模型进行训练、用定性与定量相结合的方法,选出既符合林分直径分布规律,又具有较高拟合准确度的网络模型结构为1∶2∶1,网络对象名为FRdnet2 该模型的总体累积频率拟合准确度达99.5%,径阶累积频率拟合准确度最低93%、最高99.9%、平均99%,径阶频率拟合准确度最低82%、最高99%、平均95%. 神经网络建模技术的拟合准确度好,可作为有效的林分直径分布模拟技术.      

Wireless sensor network deployment for integrating video-surveillance and data-monitoring in precision agriculture over distributed crops

Monitoring different parameters of interest in a crop has been proven as a useful tool to improve agricultural production. Crop monitoring in precision agriculture may be achieved by a multiplicity of technologies; however the use of Wireless Sensor Networks (WSNs) results in low-cost and low-power consumption deployments, therefore becoming a dominant option. It is also well-known that crops are also negatively affected by intruders (human or animals) and by insufficient control of the production process. Video-surveillance is a solution to detect and identify intruders as well as to better take care of the production process. In this paper, a new platform called Integrated WSN Solution for Precision Agriculture is proposed. The only cost-effective technology employed is IEEE 802.15.4, and it efficiently integrates crop data acquisition, data transmission to the end-user and video-surveillance tasks. This platform has been evaluated for the particular scenario of scattered crops video-surveillance by using computer simulation and analysis. The telecommunications metrics of choice are energy consumed, probability of frame collision and end-to-end latency, which have been carefully studied to offer the most appropriate wireless network operation. Wireless node prototypes providing agriculture data monitoring, motion detection, camera sensor and long distance data transmission (in the order of several kilometers) are developed. The performance evaluation of this real tests-bed scenario demonstrates the feasibility of the platform designed and confirms the simulation and analytical results.     

雨量站分布不均匀流域的降雨径流预报人工神经网络模型

为解决传统方法难以可靠预报雨量站分布不均匀流域的次降雨径流量这一水文预报难题 ,探讨了人工神经网络模型用于该类水文预报问题的可能性。实例研究表明 ,以次暴雨量及其前期影响雨量为输入、次暴雨径流总量 (净雨量 )为输出的 BP网络模型 ,预报的相对误差比蓄满产流模型预报的相对误差平均低 9.2 % ,这说明 ,人工神经网络模型可作为雨量站分布不均匀等雨量观测存在系统偏差或不足流域的降雨径流预报模型。     

基于基本竞争型神经网络的TM影像分类研究

【目的】针对传统遥感图像分类方法精度低的缺点,运用基本竞争型神经网络模型对TM影像进行分类研究。【方法】在考虑TM影像光谱信息和地表结构变化信息的基础上,应用经过基本竞争型神经网络训练后的分类器对TM影像进行分类研究,并与利用最大似然法的分类结果进行比较。【结果】研究区TM影像采用基本竞争型神经网络进行分类的总体分类精度和Kappa系数分别为89.1%和0.873,而采用最大似然法分别为70.6%和0.646,前者的分类精度明显高于后者。【结论】基本竞争型神经网络的分类结果明显优于最大似然法的分类结果。     

Soil health: model programs in the USA

Population growth, increasing drought, and natural resources degradation are significant global issues. Fortunately, management practices to improve soil health can address many of these issues in ways that are both good for the farmer and the environment. In 2012, the United States Department of Agriculture (USDA) Natural Resources Conservation Service initiated its “Unlock the Secrets in the Soil” campaign to assist farmers and ranchers with adopting soil health systems. Other notable efforts at the federal level include research and education projects by USDA Agricultural Research Service, National Institute of Food and Agriculture, and Sustainable Agriculture Research and Education program. The importance of improving soil health is also recognized far beyond federal government programs. The Soil Health Institute, a nonprofit charity, was established to safeguard and enhance the vitality and productivity of soil through scientific research and adoption. Crop commodity organizations are promoting soil health, such as the Soil Health Partnership. The Nature Conservancy is helping companies and private landowners incorporate soil health into their sustainability efforts. Such efforts are supported by the Foundation for Food and Agriculture Research, as well as by some of the leading global food companies such as General Mills. These are just a few of the many efforts to improve soil health in the USA. The significance of the expanding global population, natural resource challenges, and responsibilities to current and future generations truly make enhancing soil health a global imperative.     

武汉市春季鲜食甜玉米产量和可溶性糖含量的气象影响因子研究

为探讨气象条件对武汉市春季鲜食甜玉米产量和可溶性糖含量的影响，本文开展了分期播种试验，分析了春季鲜食甜玉米产量和可溶性糖含量随播期推移的变化，研究了不同播期处理各生育阶段的气象要素敏感性，通过数据统计、相关分析及偏相关分析等方法，分析了鲜食甜玉米产量及可溶性糖含量的气象影响因子。结果表明：春季鲜食甜玉米生育期在100 d左右，全生育期随播期的推迟而缩短；甜玉米产量与平均气温、有效积温、降水量和日照均有关系，可溶性糖含量主要与平均气温和降水量有关。     

基于T-S 模型的神经网络在小陇山森林健康评价中的应用

综合考虑T-S模型的精确性及快速性和BP神经网络的自适应能力,提出了基于T-S模型的模糊神经网络的森林健康评价模型,并利用MATLAB7.5编写程序。以小陇山油松林为例,进行计算,将结果呈送给多位小陇山林业专家,专家们对评价结果均满意。