一种基于深度可分离卷积的轻量级人体关键点检测算法

针对网络模型执行过程耗时过长的问题,受到像素切割网络架构中的编解码结构启发,设计了一种高效的轻量级主干网络,使用深度可分离卷积作为基本的卷积模块,利用了多维自学习模块(Multidimensional Self-Learning Module, MSLM)对特征矩阵进行自适应的学习来增强有用信息权重,同时使用编解码结构对其主干结构进行效率上的优化,设计出了深度可分离网络(Codec Depth Separable Network, CSDNet),相比于MobileNet性能提升了72%,精度提升了5.2%。     

基于神经网络的电子战人力资源规划分析

从现代智能化战争的电子战人力资源需求角度出发,分析了现代智能化战争中电子战人力资源规划问题,提出了基于神经网络理论的电子战人力资源规划分析思路和建议.其次,建立了兵力需求规划模型、BP神经网络和RBF神经网络预测模型以及电子战人员补充模型,将以往以定性分析为主的问题进行定量求解.最后,通过实际人员数量算例对模型进行分析...     

基于光谱技术的杨梅汁品种快速鉴别方法的研究

为了实现杨梅汁品种的快速无损鉴别,提出了一种用可见和近红外光谱技术快速鉴别杨梅汁品种的新方法。首先采用偏最小二乘法进行模式特征分析,经过交互验证法判别,确定最佳主成分数为9。完成特征提取后,将这9个主成分作为神经网络的输入变量,建立了三层BP神经网络,实现类别预测的同时也完成了数学建模与优化分析工作。3个品种的杨梅汁样本数均为20,共计60个样本。在神经网络学习中,将其分成训练集样本51个和预测集样本9个。对9个未知样本进行预测,准确率为100％。说明本文提出的基于光谱技术和模式识别的方法具有很好的分类和鉴别能力。     

Special issue on system-integrated intelligence. New challenges for product and production engineering

The ubiquitous development of information and communication technology enables new opportunities for products as well as for production and manufacturing systems. These systems will be able to learn and adapt their behaviour during the systems operation for a continuos optimization. This results in an increasing structural complexity and dynamics of products, production networks, processes and organizations, which in turn requires an on-going adaption and reinvention of the organizing principles and solutions. Therefore, new products as well as their corresponding production and logistic processes spawn research activities in the field of advanced information techniques and system integrated intelligence to cope with the complexity and dynamics of future manufacturing networks.     

浅析微课在《网络创业》课程中的应用

近年来,国内的微课研究热潮来势凶猛.本文以网络创业课程为研究对象,研究了微课在高职教育课程中的应用,同时根据授课情况验证了微课模式的可行性,并取得了预期的效果.     

Improvement of permanent memory effect in PDLC films using TX-100 as an additive

A number of polymer dispersed liquid crystal (PDLC) films based on dimethacrylate monomers were synthesised by polymer-induced phase separation (PIPS) using polyethylene glycol p-(1,1,3,3-tetramethylbutyl)-phenyl ether (TX-100) as an additive up to total solution weight of 10% aiming to modify the initial anchorage force of the liquid crystal (LC) molecules to the polymeric matrix. The effects of the addition of this additive to the PDLC films were studied resorting to electro-optical studies. Using a fitting model, several parameters were correlated such as the permanent memory effect (PME), the voltage required to achieve 90% of maximum transmittance (E90), the average elastic constant (K) and the rotational viscosity of the director (γ). The use of TX-100 results on an increase on the PME and γ, and decrease on the E90 and K.     

Review and perspective on encrypted traffic identification research

Considering the importance of encrypted traffic identification technology and existing research work,first,the type of encrypted traffic identification according to the demand of traffic analysis were introduced,such as protocols,applications and services.Second,the encrypted traffic identification technology was summarized,and identification technology was compared from multiple views.Third,the deficiencies and the affecting factors of the existing encrypted traffic identification technologies were induced,such as tunneling,traffic camouflage technology,new protocols of HTTP/2.0 and QUIC.Finally,prospect trends and directions of future research on encrypted traffic identification were discussed.     

基于物联网无线通信的智能感知校园服务平台

以物联网技术为核心,整合目前校内分离的各个信息化系统,建立校内资源共享的综合信息门户。运用RFID技术、传感网技术、云计算等多种先进技术建设基于物联网理念的感知校园服务平台。     

The application of Bayesian network in Uncertainty management

Uncertainty management is playing an important role in many fields,especially in AI(artificial intelligence).This paper introduces the concept of probability and the Bayesian Network which are widely used in uncertainty management field.Additionally,two examples are completed and analyzed in the paper.     

k-Edge Failure Resilient Network Design

We design a network that supports a feasible multicommodity flow even after the failures of any k edges. We present a mixed-integer linear program (MILP), a cutting plane algorithm, and a column-and-cut algorithm. The algorithms add constraints to repair vulnerabilities in partial network designs. Empirical studies on previously unsolved instances of SNDlib demonstrate their effectiveness.