基于M-DeepLab网络的速度建模技术研究

本文提出了一种适用于速度建模方法的M-DeepLab网络框架,该网络将地震炮集记录作为输入,网络主体使用轻量级MobileNet,以此提升网络训练速度;并在编码环节ASPP模块后添加了Attention模块,且在解码环节将不同网络深度的速度特征进行了融合,既获得了更多的速度特征,又保留了网络浅部的速度信息,防止出现网络退化和过拟合问题。模型测试证明,M-DeepLab网络能够实现智能、精确的速度建模,简单模型、复杂模型以及含有噪声数据复杂模型的智能速度建模,均取得了良好的效果。相较DeepLabV3+网络,本文方法对于速度模型界面处的预测,特别是速度突变区域的预测,具有更高的预测精度,从而验证了该方法精确性、高效性、实用性和抗噪性。

基于m×2正则化交叉验证的神经网络超参数调优方法

超参数调优是神经网络建模的关键问题。针对传统的超参数调优方法存在的问题,该文提出了一种基于m×2正则化交叉验证的超参数调优方法。目的是给出一种适用于复杂模型、大数据集背景下的计算开销较小且稳健的超参数调优方法。该方法的思想是从完整的数据集上选取少部分数据进行调优,避免模型在数据集较大时非常耗时的超参数调优难题;在m×2交叉验证的基础上设置正则化条件均衡训练集与验证集之间的分布差异,从而减少分布不一致带来的性能波动;使用信噪比作为调优的优化目标,从而可以综合考虑模型性能评价指标的均值和方差;并采用正交设计选择相关性较低的超参数组合以提高调优效率。以命名实体任务为例进行实验,在CoNLL 2003数据集上的实验结果显示,提出的调优方法能够选到和网格搜索性能上没有显著差异的超参数组合,且调优时间可显著降低约66%。

m^(6)A相关基因在激素性股骨头坏死中的生物信息学分析

背景:m^(6)A修饰与股骨头坏死的发生发展相关,但在激素性股骨头坏死中的作用尚不清楚。目的:基于GEO数据库,采用生物信息学方法分析激素性股骨头坏死中表达差异的m^(6)A基因及互作miRNAs,探寻其潜在发病机制。方法:在GEO数据库中检索并下载与激素性股骨头坏死相关的mRNA表达谱数据集(GSE123568),通过R软件对数据集进行差异基因筛选及GO功能、KEGG通路富集分析。识别差异基因中的m^(6)A差异表达基因(m^(6)A-DEGs)并对其进行GO功能与KEGG通路富集分析,比较m^(6)A-DEGs的表达量并分析它们之间的相关性。最后通过Cytoscape构建m^(6)A-DEGs的PPI互作网络及筛选核心基因。使用TargetScan,miRTarBase和miRBD数据库预测m^(6)A-DEGs相关的潜在miRNAs,同时使用ChIPBase及hTFtarget数据库预测7个核心基因潜在转录因子,然后分别构建m^(6)A-miRNA与转录因子m^(6)A调控网络。最后使用数据集GSE74089验证7个核心m^(6)A-DEGs的表达水平。结果与结论:①从数据集中共筛选出2460个差异表达的基因,其中1455个上调,1005个下调。②从数据集中筛选出了14个m^(6)A-DEGs,包括3个下调和11个上调基因,m^(6)A-DEGs在激素性股骨头坏死中的表达具有显著差异(P<0.05),Spearman分析表明它们之间具有一定相关性。③m^(6)A-DEGs的GO和KEGG富集分析主要集中在骨髓细胞分化与发育、免疫受体与细胞因子受体活性、破骨细胞分化、AMPK与白细胞介素17信号通路。④m^(6)A-DEGs前7个核心基因包括YTHDF3,YTHDF1,YTHDF2,ALKBH5,METTL3,HNRNPA2B1及HNRNPC,它们在miRTarBase,miRDB和TargetScan数据库中共有44个miRNA重叠,在ChIPBase及hTFtarget数据库中共有79个重叠转录因子。⑤在GSE74089数据集中有6个核心m^(6)A-DEGs的表达水平与GSE123568数据集一致。⑥结果证实,根据生物信息学方法筛选的7个m^(6)A-DEGs可能通过调控多个miRNA、转录因子和AMPK及白细胞介素17信号通路表达,进而影响激素性股骨头坏死中骨髓细胞分化发育与破骨细胞分化,为进一步深入研究激素性股骨头坏死的发病机制和靶向治疗提供了数据支持和研究方向。

基于VMD-DBO-LSTM的空气质量预测

针对传统空气质量预测模型收敛速度慢,精度低的问题,提出一种基于变分模态分解(variational mode decomposition,VMD)和蜣螂优化算法(dung beetle optimizer,DBO)优化长短期记忆网络(long short term memory,LSTM)的预测模型。首先,针对AQI原始数据具有大量噪声的问题,使用VMD方法对非平稳信号进行模态分解以降低噪声对预测结果的影响从而获得多个不同特征的模态分量;其次,针对LSTM靠人工经验调参存在一定局限性,利用DBO算法对LSTM模型参数进行优化;最后,对分解后的各个子序列使用LSTM模型预测,将各个子序列进行叠加得到最后的预测结果。实验结果表明,VMD对非平稳数据的分解有助于提高预测精度,VMD-DBO-LSTM模型的性能较其他模型均有不同程度的提高,该模型预测的均方根误差为4.73μg/m^(3),平均绝对误差为3.61μg/m^(3),拟合度达到了97.8%。

An Auto-configuration 4M Group Management using Wireless Sensor Networks for Reconfigurable Manufacturing System

The first tier of automotive manufacturers has faced to pressures about move,modify,updating tasks for manufacturing resources in production processes from demand response of production order sequence for motor company and process innovation purpose for productivity. For meets this requirements,it has to require absolutely lead time to re-wiring of physical interface for production equipment,needs for change existing program and test over again.For prepare this constraints,it needs studying an auto-configuration functions that build for both visibility and flexibility based on the 4M(Man,Machine,Material, Method)group management which is supports from WSN (Wireless Sensor Network)of the open embedded device called M2M(Machine to Machine)and major functions of middleware including point manager for real-time device communication,real-time data management,Standard API (Application Program Interface)and application template management.To be application system to RMS (Reconfigurable Manufacturing System)for rapidly response from various orders and model from motor company that is beginning to establishing the mapping of manufacturing resources of 4M using WSN.

An Auto-Configuration of 4M Group Management Using Wireless Sensor Networks

The first tier of automotive manufacturers has faced to pressures about move, modify, updating tasks for manufacturing resources in production processes from demand response of production order sequence for motor company and process innovation purpose for productivity. For meets this requirements, it has to require absolutely lead time to re-wiring of physical interface for production equipment, needs for change existing program and test over again. For prepare this constraints, it needs studying an auto-configuration functions that build for both visibility and flexibility based on the 4M (Man, Machine, Material, Method) group management which is supports from WSN (Wireless Sensor Network) of the open embedded device called M2M (Machine to Machine) and major functions of middleware including point manager for real-time device communication, real-time data management, Standard API (Application Program Interface) and application template management. To be application system to RMS (Reconfigurable Manufacturing System) for rapidly response from various orders and model from motor company that is beginning to establishing the mapping of manufacturing resources of 4M using WSN.

工控网络Modbus TCP协议异常报文检测方法

工业控制网络系统包含Modbus报文传输数据、TCP/IP主机通信行为数据、工控协议数据等大量复杂数据,容易造成数据冗余,降低异常报文检测精度,增加检测时间,从而增加工控网络被攻击几率。为此,提出一种基于支持向量机(SVM)的工控网络Modbus TCP协议异常报文检测方法。利用SVM建立工控网络分层数据模型,在应用层中定义Modbus协议报文,使用核函数对非平稳的随机Modbus协议报文进行离散处理,减少工控网络数据冗余。结合m阶马尔科夫序列,在TCP/IP数据链路层中对离散Modbus协议报文进行异常特征提取,完成Modbus报文异常检测。仿真测试结果表明:所提方法可以有效降低数据冗余,工控网络重复数据存储率为0.39%;Modbus TCP协议异常报文检测精度为95.73%;Modbus TCP协议异常报文检测时间为0.5 ms,以期为工控网络安全提供技术支持。

LTE-M技术在大连地铁13号线信号CBTC系统的规划研究

随着通信技术的迭代发展,基于CBTC的无线列车控制系统已逐步代替基于轨道电路的车地通信技术。其中,WLAN技术应用广泛。大连城市轨道交通中地铁1、2号线信号数据通信系统采用2.4 GHz频段的WLAN技术。近几年,随着我国无线网络的迅速普及、Wi-Fi的大量使用,对地铁1、2号线信号系统车地无线通信的稳定性带来了隐患。因此,大连地铁13号线在修建过程中,信号数据通信网络DCS系统采用了1.8 GHz专用频段的LTE-M技术。文章以此为切入点,对大连地铁13号线信号系统LTE-M技术方案进行详细介绍,同时通过对WLAN技术以及LTE技术对比分析,阐述LTE技术在轨道交通车地无线通信系统中具有的技术优势。

基于OSU的M-OTN组网模式与现网试验

基于光业务单元(optical service unit,OSU)的面向城域优化的光传送网络(metro-optimized OTN,M-OTN)创新技术方案可以提供低时延和带宽灵活调整的高品质专线业务,提升光传送网络(optical transport network,OTN)对小颗粒专线业务的承载能力和效率。重点探讨了OSU技术在OTN中的应用模式,包括OSU在现网中的应用价值、叠加层(Overlay)与底层(Underlay)组网模式比较、OSU到光数据单元k(optical data unit-k,ODUk)复用方式与多业务统一承载等几个方面。通过现网试验对OSU技术的业务能力和长期运行可靠性进行了充分验证,为后续OSU组网模式的选择提供技术依据。

Diagnosabilities of exchanged hypercube networks under the pessimistic one-step diagnosis strategy

The exchanged hypercube EH（s, t） （where s ≥ 1 and t ≥ 1） is obtained by systematically reducing links from a regular hypercube Q,＋t＋l. One-step diagnosis of exchanged hypercubes which involves only one testing phase during which processors test each other is discussed. The diagnosabilities of exchanged hypercubes are studied by using the pessimistic one-step diagno- sis strategy under two kinds of diagnosis models： the PMC model and the MM model. The main results presented here are the two proofs that the degree of diagnosability of the EH（s, t） under pessimistic one-step tl/tl fault diagnosis strategy is 2s where I ≤ s ≤ t （respectively, 2t, where 1 ≤ t ≤ s） based on the PMC model and that it is also 2s where 1 ≤ s ≤ t （respectively, 2t, where 1 ≤ t ≤ s） based on the MM＊ model.

Energy-Efficient User Association in Heterogeneous Networks with M2M/H2H Coexistence under QoS Guarantees

In this paper, we propose an energy efficient user association scheme for uplink heterogeneous networks with machine-to-machine(M2M) and human-to-human(H2H) coexistence. A group based random access protocol is designed for massive number of machine-typecommunications(MTC) user equipments'(UEs) transmissions. A user association problem for UEs' energy efficiency maximization is formulated considering the HTC UEs' quality of service(QoS) guarantees and load balance among multiple BSs, simultaneously. A distributed iterative algorithm is developed to solve the optimization problem. In addition, the convergence of the proposed algorithm is proved. Simulation results show that our proposed scheme outperforms other schemes in terms of energy efficiency and QoS guarantees.

Rate Allocation Scheme for Machine-to-Machine Service Based on 3GPP in Heterogeneous Wireless Networks

In this study,an explicit adaptive traffic allocation scheme for Machine-to-Machine(M2M)service is proposed to achieve optimum distribution in heterogeneous networks.Based on the characteristics of M2M services,the presented scheme is formulated as a convex optimization problem that maximises the utility of the M2M service,and then determines how to allocate the total rate among the multiple access networks.The analysis and numerical simulations indicate that the proposed scheme makes a significant improvement in performance compared with the traditional schemes.

McCulloch-Pitts人工神经元模型80周年纪念:思想、方法与意义

1943年,Warren McCulloch和Walter Pitts发表论文《神经活动内在思想的逻辑演算》,证明可以使用逻辑演算描述神经网络的运行机理,拓展了计算神经科学的领域前沿,奠定了人工神经网络的发展基础。恰逢M-P论文发表80周年之际,结合时代背景和作者学术生涯的变迁,探索M-P理论的思想源泉;借助原文中的示例,概述M-P模型的基本原理;以作者对循环因果和反馈机制的研究为主线,探讨M-P模型对信息科学发展的影响、对控制论发展的贡献,及其在当今人工智能大模型时代的意义。

Delay optimization for planar wireless sensor network with N-policy

In N-policy, the nodes attempt to seize the channel when the number of packets in the buffer approaches N. The performance of N-policy on the energy efficiency is widely studied in the past years. And it is presented that there exists one optimal N to minimize the energy consumption. However, it is noticed that the delay raised by N-policy receives little attention. This work mathematically proves the delay to monotonically increase with increasing N in the collision-unfree channel. For planar network where the near-to-sink nodes burden heavier traffic than the external ones, the data stemming from the latter undergo longer delay.The various-N algorithm is proposed to address this phenomenon by decreasing the threshold N of outer nodes. Without the impacting on the network longevity, the maximum delay among the network has decreased 62.9% by the algorithm. Extensive simulations are given to verify the effectiveness and correctness of our analysis.

MobiHealthcare System: Body Sensor Network Based M-Health System for Healthcare Application

M-health, which is known as the practice of medical and public health supported by mobile devices such as mobile phones and PDAs for delivering medical and healthcare services, is currently being heavily developed to keep pace with the continuously rising demand for personalized healthcare. To this end, the MobiHealthcare system, which provides a personalized healthcare based on body sensor network, is developed. The system includes various body sensors to collect physiological signals specifically for different requirements, a cell phone to facilitate the joint processing of spatially and temporally collected medical data from different parts of the body for resource optimization and systematic health monitoring, a server cluster with great data storage capacity, powerful analysis capabilities to provide data storage, data mining and visualization. Compared with existing M-Health system, the MobiHealthcare system is characteristics of low coupling and powerful parallel computing capabilities. Various healthcare applications have been implemented in the proposed system to demonstrate its effectiveness in providing a powerful platform.

GLOBAL STABILITY IN HOPFIELD NEURAL NETWORKS WITH DISTRIBUTED TIME DELAYS

In this paper, without assuming the boundedness, monotonicity and differentiability of the activation functions, the conditions ensuring existence, uniqueness, and global asymptotical stability of the equilibrium point of Hopfield neural network models with distributed time delays are studied. Using M-matrix theory and constructing proper Liapunov functionals, the sufficient conditions for global asymptotic stability are obtained.

A NEW METHOD FOR SOLVING MSDE BASED ON WAVELET NEURAL NETWORKS

In this paper, a new method to solve multiscale difference equation(MSDE) with the M-band wavelet neural networks is proposed. It is shown that the method has many advantages over the existing methods and enlarges the range of the solvable equations.

Deployment of Polar Codes for Mission-Critical Machine-Type Communication Over Wireless Networks

Mission critical Machine-type Communication(mcMTC),also referred to as Ultra-reliable Low Latency Communication(URLLC),has become a research hotspot.It is primarily characterized by communication that provides ultra-high reliability and very low latency to concurrently transmit short commands to a massive number of connected devices.While the reduction in physical(PHY)layer overhead and improvement in channel coding techniques are pivotal in reducing latency and improving reliability,the current wireless standards dedicated to support mcMTC rely heavily on adopting the bottom layers of general-purpose wireless standards and customizing only the upper layers.The mcMTC has a significant technical impact on the design of all layers of the communication protocol stack.In this paper,an innovative bottom-up approach has been proposed for mcMTC applications through PHY layer targeted at improving the transmission reliability by implementing ultra-reliable channel coding scheme in the PHY layer of IEEE 802.11a standard bearing in mind short packet transmission system.To achieve this aim,we analyzed and compared the channel coding performance of convolutional codes(CCs),low-density parity-check(LDPC)codes,and polar codes in wireless network on the condition of short data packet transmission.The Viterbi decoding algorithm(VA),logarithmic belief propagation(Log-BP)algorithm,and cyclic redundancy check(CRC)successive cancellation list(SCL)(CRC-SCL)decoding algorithm were adopted to CC,LDPC codes,and polar codes,respectively.Consequently,a new PHY layer for mcMTC has been proposed.The reliability of the proposed approach has been validated by simulation in terms of Bit error rate(BER)and packet error rate(PER)vs.signal-to-noise ratio(SNR).The simulation results demonstrate that the reliability of IEEE 802.11a standard has been significantly improved to be at PER=10−5 or even better with the implementation of polar codes.The results also show that the general-purpose wireless networks are prominent inproviding short packet mcMTC with the modification needed.

Practical Timing Synchronization for Network Dynamics in Large Machine-to-Machine Networks

Efficient multi-machine cooperation and network dynamics still remain open that jeopardize great applications in largescale machine-to-machine(M2M) networks. Among all possible machine cooperation controls, to synchronize tremendous machines in a timing-efficient brings one of the greatest challenge and serves as the foundation for any other network control policies. In this paper, we propose a linear-time synchronization protocol in large M2M networks. Specifically, a closed-form of synchronization rate is provided by developing the statistical bounds of the second smallest eigenvalue of the graph Laplacian matrix. These bounds enable the efficient control of network dynamics, facilitating the timing synchronization in networks. Through a practical study in Metropolis, simulation results confirm our theoretical analysis and provide effective selection of wireless technologies, including Zigbee, Wi-Fi, and cellular systems, with respect to the deployed density of machines. Therefore, this paper successfully demonstrates a practical timing synchronization, to make a breakthrough of network dynamic control in real-world machine systems, such as Internet of Things.

Soliton interactions and asymptotic state analysis in a discrete nonlocal nonlinear self-dual network equation of reverse-space type

We propose a reverse-space nonlocal nonlinear self-dual network equation under special symmetry reduction,which may have potential applications in electric circuits.Nonlocal infinitely many conservation laws are constructed based on its Lax pair.Nonlocal discrete generalized(m,N−m)-fold Darboux transformation is extended and applied to solve this system.As an application of the method,we obtain multi-soliton solutions in zero seed background via the nonlocal discrete N-fold Darboux transformation and rational solutions from nonzero-seed background via the nonlocal discrete generalized(1,N−1)-fold Darboux transformation,respectively.By using the asymptotic and graphic analysis,structures of one-,two-,three-and four-soliton solutions are shown and discussed graphically.We find that single component field in this nonlocal system displays unstable soliton structure whereas the combined potential terms exhibit stable soliton structures.It is shown that the soliton structures are quite different between discrete local and nonlocal systems.Results given in this paper may be helpful for understanding the electrical signals propagation.