Insights into microbiota community dynamics and flavor development mechanism during golden pomfret(Trachinotus ovatus)fermentation based on single-molecule real-time sequencing and molecular networking analysis

Popular fermented golden pomfret(Trachinotus ovatus)is prepared via spontaneous fermentation;however,the mechanisms underlying the regulation of its flavor development remain unclear.This study shows the roles of the complex microbiota and the dynamic changes in microbial community and flavor compounds during fish fermentation.Single-molecule real-time sequencing and molecular networking analysis revealed the correlations among different microbial genera and the relationships between microbial taxa and volatile compounds.Mechanisms underlying flavor development were also elucidated via KEGG based functional annotations.Clostridium,Shewanella,and Staphylococcus were the dominant microbial genera.Forty-nine volatile compounds were detected in the fermented fish samples,with thirteen identified as characteristic volatile compounds(ROAV>1).Volatile profiles resulted from the interactions among the microorganisms and derived enzymes,with the main metabolic pathways being amino acid biosynthesis/metabolism,carbon metabolism,and glycolysis/gluconeogenesis.This study demonstrated the approaches for distinguishing key microbiota associated with volatile compounds and monitoring the industrial production of high-quality fermented fish products.

Threshold-Based Software-Defined Networking(SDN)Solution for Healthcare Systems against Intrusion Attacks

The healthcare sector holds valuable and sensitive data.The amount of this data and the need to handle,exchange,and protect it,has been increasing at a fast pace.Due to their nature,software-defined networks(SDNs)are widely used in healthcare systems,as they ensure effective resource utilization,safety,great network management,and monitoring.In this sector,due to the value of thedata,SDNs faceamajor challengeposed byawide range of attacks,such as distributed denial of service(DDoS)and probe attacks.These attacks reduce network performance,causing the degradation of different key performance indicators(KPIs)or,in the worst cases,a network failure which can threaten human lives.This can be significant,especially with the current expansion of portable healthcare that supports mobile and wireless devices for what is called mobile health,or m-health.In this study,we examine the effectiveness of using SDNs for defense against DDoS,as well as their effects on different network KPIs under various scenarios.We propose a threshold-based DDoS classifier(TBDC)technique to classify DDoS attacks in healthcare SDNs,aiming to block traffic considered a hazard in the form of a DDoS attack.We then evaluate the accuracy and performance of the proposed TBDC approach.Our technique shows outstanding performance,increasing the mean throughput by 190.3%,reducing the mean delay by 95%,and reducing packet loss by 99.7%relative to normal,with DDoS attack traffic.

The Caching and Pricing Strategy for Information-Centric Networking with Advertisers’Participation

As users’access to the network has evolved into the acquisition of mass contents instead of IP addresses,the IP network architecture based on end-to-end communication cannot meet users’needs.Therefore,the Information-Centric Networking(ICN)came into being.From a technical point of view,ICN is a promising future network architecture.Researching and customizing a reasonable pricing mechanism plays a positive role in promoting the deployment of ICN.The current research on ICN pricing mechanism is focused on paid content.Therefore,we study an ICN pricing model for free content,which uses game theory based on Nash equilibrium to analysis.In this work,advertisers are considered,and an advertiser model is established to describe the economic interaction between advertisers and ICN entities.This solution can formulate the best pricing strategy for all ICN entities and maximize the benefits of each entity.Our extensive analysis and numerical results show that the proposed pricing framework is significantly better than existing solutions when it comes to free content.

Open-Source Software Defined Networking Controllers:State-of-the-Art,Challenges and Solutions for Future Network Providers

Software Defined Networking(SDN)is programmable by separation of forwarding control through the centralization of the controller.The controller plays the role of the‘brain’that dictates the intelligent part of SDN technology.Various versions of SDN controllers exist as a response to the diverse demands and functions expected of them.There are several SDN controllers available in the open market besides a large number of commercial controllers;some are developed tomeet carrier-grade service levels and one of the recent trends in open-source SDN controllers is the Open Network Operating System(ONOS).This paper presents a comparative study between open source SDN controllers,which are known as Network Controller Platform(NOX),Python-based Network Controller(POX),component-based SDN framework(Ryu),Java-based OpenFlow controller(Floodlight),OpenDayLight(ODL)and ONOS.The discussion is further extended into ONOS architecture,as well as,the evolution of ONOS controllers.This article will review use cases based on ONOS controllers in several application deployments.Moreover,the opportunities and challenges of open source SDN controllers will be discussed,exploring carriergrade ONOS for future real-world deployments,ONOS unique features and identifying the suitable choice of SDN controller for service providers.In addition,we attempt to provide answers to several critical questions relating to the implications of the open-source nature of SDN controllers regarding vendor lock-in,interoperability,and standards compliance,Similarly,real-world use cases of organizations using open-source SDN are highlighted and how the open-source community contributes to the development of SDN controllers.Furthermore,challenges faced by open-source projects,and considerations when choosing an open-source SDN controller are underscored.Then the role of Artificial Intelligence(AI)and Machine Learning(ML)in the evolution of open-source SDN controllers in light of recent research is indicated.In addition,the challenges and limitations associated with deploying open-source SDN controllers in production networks,how can they be mitigated,and finally how opensource SDN controllers handle network security and ensure that network configurations and policies are robust and resilient are presented.Potential opportunities and challenges for future Open SDN deployment are outlined to conclude the article.

Information Centric Networking Based Cooperative Caching Framework for 5G Communication Systems

The demands on conventional communication networks are increasing rapidly because of the exponential expansion of connected multimedia content.In light of the data-centric aspect of contemporary communication,the information-centric network(ICN)paradigm offers hope for a solution by emphasizing content retrieval by name instead of location.If 5G networks are to meet the expected data demand surge from expanded connectivity and Internet of Things(IoT)devices,then effective caching solutions will be required tomaximize network throughput andminimize the use of resources.Hence,an ICN-based Cooperative Caching(ICN-CoC)technique has been used to select a cache by considering cache position,content attractiveness,and rate prediction.The findings show that utilizing our suggested approach improves caching regarding the Cache Hit Ratio(CHR)of 84.3%,Average Hop Minimization Ratio(AHMR)of 89.5%,and Mean Access Latency(MAL)of 0.4 s.Within a framework,it suggests improved caching strategies to handle the difficulty of effectively controlling data consumption in 5G networks.These improvements aim to make the network run more smoothly by enhancing content delivery,decreasing latency,and relieving congestion.By improving 5G communication systems’capacity tomanage the demands faced by modern data-centric applications,the research ultimately aids in advancement.

Networking Observation and Applications of Chinese Ocean Satellites

This paper presents the networking observation capabilities of Chinese ocean satellites and their diverse applications in ocean disaster prevention,ecological monitoring,and resource development.Since the inaugural launch in 2002,China has achieved substantial advancements in ocean satellite technology,forming an observation system composed of the HY-1,HY-2,and HY-3 series satellites.These satellites are integral to global ocean environmental monitoring due to their high resolution,extensive coverage,and frequent observations.Looking forward,China aims to further enhance and expand its ocean satellite capabilities through ongoing projects to support global environmental protection and sustainable development.

结合太阳辐射量计算与CNN-LSTM组合的光伏功率预测方法研究

为了提高模型预测性能,提出一种综合太阳辐射模型及深度学习的光伏功率预测模型。首先,利用太阳辐射机理建立太阳辐射模型(SRM),估算出水平面上总辐射值,再由斜面辐照度转换方法计算出光伏组件所接收的斜面辐射值。其次,通过皮尔逊相关分析法筛选出对光伏功率影响较大的主要因素,将斜面辐射计算值及主要影响因素作为输入,采用卷积神经网络(CNN)和长短期记忆网络(LSTM)建立光伏功率SRM-CNN-LSTM预测模型。分别利用春夏秋冬四季典型日的数据开展对比实验,结果表明:与几种其他方法相比,该文方法具有更好的预测效果。

基于混合分解和PCG-BiLSTM的风速短期预测

为降低风速的随机性对风力发电的影响,提高风速短期预测的精准度,提出一种基于混合分解、双通道输入、多分支PCG-BiLSTM深度学习模型的短期风速预测方法。首先,将全年风速数据分为春、夏、秋、冬4个季度,选取春季作为主要实验对象;其次,利用奇异谱分解(SSD)和变分模态分解(VMD)以降低原始春季风速数据复杂度,生成具有不同模态且复杂度低的子分量,两种不同模式子分量组合为混合分量,实现不同模式分解算法的优势互补;最后,将混合分量以双通道的形式输入到多分支PCG-BiLSTM深度学习模型中,其模型的每个分支由卷积神经网络(CNN)与门控循环单元(GRU)并联组成时空特征提取模块,用于提取两种分解分量组合的混合分量的时空特征,各分支提取对应混合分量的时空特征经聚合后再由双向长短期记忆网络(BiLSTM)进一步提取风速信号的正向和反向双向波动规律,进而得到最终的风速预测结果。多组实验结果表明:提出的组合预测方法在短期风速预测中具有较高的精度和泛化能力,优于其他传统预测方法。

基于CNN-LSTM的大坝变形组合预测模型研究

为了提高大坝变形预测模型精度和泛化能力,建立了一种基于卷积神经网络(Convolutional neural networks,CNN)与深度学习长短期记忆(Long short-term memory,LSTM)神经网络的组合预测模型CNN-LSTM。该模型先利用CNN提取大坝变形监测时间序列的特征,再利用LSTM生成特征描述,该模型精度高、泛化能力强。以柏叶口水库混凝土面板堆石坝为例,经过CNN-LSTM模型计算,将模型变形预测值与原型监测资料进行对比,再与LSTM模型及CNN模型的预测结果进行对比。结果表明,CNN-LSTM模型预测值最接近监测资料实测结果。

基于改进StackCNN网络和集成学习的脑电信号视觉分类算法

针对直接使用图像诱发的脑电信号进行视觉分类的现有研究少,并且视觉分类的平均准确率低等问题,设计了一种卷积神经网络(CNN)和集成学习相结合的方法,用于学习脑电信号相关的视觉特征表示。通过在StackCNN网络中加入K-max池化方法,解决在提取脑电特征时信息丢失的问题,并结合Bagging算法增强网络的泛化能力,该方法称为StackCNN-B。采用基于残差神经网络(ResNet)回归对图像进行分类,验证StackCNN-B方法在图像分类上的性能。消融实验及与现有研究对比实验的结果表明:所提方法识别准确率较高,在学习脑电信号的视觉特征表示上的平均准确率达到99.78%,在图像分类上的平均准确率达到96.45%,与Bi-LSTM-AttGW方法相比,平均提高了0.28百分点和2.97百分点。研究结果验证了脑电信号可以有效地解码与视觉识别相关的人类大脑活动,也表明所提出StackCNN-B模型的优越性。

基于Transformer和自适应特征融合的矿井低照度图像亮度提升和细节增强方法

高质量矿井影像为矿山安全生产提供保障,也有利于提高后续图像分析技术的性能。矿井影像受低照度环境的影响,易出现亮度低,照度不均,颜色失真,细节信息丢失严重等问题。针对上述问题,提出一种基于Transformer和自适应特征融合的矿井低照度图像亮度提升和细节增强方法。基于生成对抗思想搭建生成对抗式主体模型框架,使用目标图像域而非单一参考图像驱动判别器监督生成器的训练,实现对低照度图像的充分增强;基于特征表示学习理论搭建特征编码器,将图像解耦为亮度分量和反射分量,避免图像增强过程中亮度与颜色特征相互影响从而导致颜色失真问题;设计CEM-Transformer Encoder通过捕获全局上下文关系和提取局部区域特征,能够充分提升整体图像亮度并消除局部区域照度不均;在反射分量增强过程中,使用结合CEM-Cross-Transformer Encoder的跳跃连接将低级特征与深层网络处特征进行自适应融合,能够有效避免细节特征丢失,并在编码网络中添加ECA-Net,提高浅层网络的特征提取效率。制作矿井低照度图像数据集为矿井低照度图像增强任务提供数据资源。试验显示,在矿井低照度图像数据集和公共数据集中,与5种先进的低照度图像增强算法相比,该算法增强图像的质量指标PSNR、SSIM、VIF平均提高了16.564%,10.998%,16.226%和14.438%,10.888%,14.948%,证明该算法能够有效提升整体图像亮度,消除照度不均,避免颜色失真和细节丢失,实现矿井低照度图像增强。

适用于区域建筑群实时震害模拟的LSTM-FC组合深度网络模型研究

建筑物破坏在地震灾害中往往会导致巨大损失,对城市建筑群进行灾前灾时的震害预测具有重要意义。传统BP(back propagation)网络和CNN(convolutional neural networks)网络等人工智能方法在进行震害预测时多集中于提取建筑物信息。然而,这些方法在处理地震波的时序数据方面有所不足,导致其在整合和分析对地震灾害预测至关重要的时序相关因素时效果有限。因此,本文提出一种耦合LSTM(long short-term memory)和FC(fully connected)神经网络的震害预测方法。LSTM网络擅长处理具有时间序列特性的地震波信息,能够捕捉和分析随时间变化的地震波动模式。同时,全连接网络可用于综合分析所有相关的震害因子。通过对云浮地区265栋典型钢混建筑进行指标量化并确定输入指标(震害影响因子)和输出指标(震害指数),利用LSTM-FC组合深度网络、CNN网络和BP网络模型对数据进行训练并优化。通过将LSTM-FC网络模型的预测结果与弹塑性时程分析比较,发现该模型在拟合效果和精度方面优于传统的BP和CNN模型。拟合效果提升了36.8%和10.6%,精度分别提升了77.6%和91.7%,表明LSTM-FC网络在地震损害预测上更为有效。同时,将该方法应用于广东省云浮市钢混结构群震害预测,构建的易损性矩阵与华南地区的易损性矩阵均值进行了对比,显示误差相对较小,说明该模型不仅理论上可行,在实际应用中也能表现出较高的准确性和有效性。

基于CNN-LSTM的机床滚动轴承性能退化趋势和寿命预测

滚动轴承作为机床主轴的关键部件,其剩余寿命预测直接决定着整机设备的剩余寿命。若不能及时地预知滚动轴承的健康状态或损伤情况,不仅会影响维修策略的制定,还会造成级联故障,易造成机床灾难性的事故。针对大数据下滚动轴承振动信号的自适应故障特征提取和智能诊断问题,构建卷积神经网络和长短期记忆网络(CNN-LSTM)相结合的寿命预测模型,它可以避免人工参与的影响,实现网络的互补优势。对滚动轴承的退化状态以及剩余寿命进行预测,并与卷积神经网络(CNN)、长短时记忆神经网络(LSTM)进行对比实验。结果表明:所提方法CNN-LSTM有着较高的预测准确度。

基于Partial New Causality的因果脑网络情绪识别

为了研究情绪产生过程中脑区以及通道之间的因果作用,在部分格兰杰与新型因果关系的基础上,提出一种用于研究时间序列之间因果关系的部分新型因果关系(PNC)方法。在不同情绪下选取脑区内的8个通道,用PNC计算脑区内通道之间的因果连接关系,根据连接关系构建因果网络;对因果网络中节点的信息流向和介数属性进行分析,将PNC因果网络和Granger因果网络节点之间的因果连接视为一种特征送入SVM中训练分类。实验结果表明,基于PNC因果网络和Granger因果网络的平均识别精度分别为76.4%和68.5%,PNC可用于计算时间序列之间的因果关系。

一种融合GA和LSTM的边坡变形预测优化网络模型及其应用

考虑到BP神经网络模型忽略边坡监测数据存在的时间相关性,以及LSTM模型由于超参数选择存在主观性而易陷入局部最优等问题,提出一种基于遗传算法和长短期记忆网络(GA-LSTM)相结合的边坡变形预测模型,以发挥遗传算法全局搜索能力和LSTM预测时序数据的优势。以海明矿业露天采场边坡为研究对象,分别采用BP神经网络模型、LSTM网络模型以及GA-LSTM网络模型对边坡监测点GNSS49变形进行预测分析,并对比各模型达到收敛条件的时间。结果表明,GA-LSTM模型与其他模型达到同一收敛条件的时间差异不大,GA-LSTM模型的拟合准确度在0.1~0.2 mm,是LSTM神经网络模型的5~7倍,是BP神经网络模型的10~20倍,具有较高的精度和稳定性,其预测值与实际监测数据基本一致,可为矿山边坡的安全生产、管理以及决策控制提供科学依据。

基于scSE非局部双流ResNet网络的行为识别

针对双流网络对包含冗余信息的视频帧存在识别率低的问题,在双流网络的基础上引入scSE(Spatial and Channel Squeeze&Excitation Block)和非局部操作,构建SC_NLResNet行为识别框架。该框架将视频划分为等分不重叠的时序段并在每段上稀疏采样,提取RGB帧以及光流图作为scSE模块的输入;将经过scSE处理的特征输入非局部双流ResNet网络中,融合各分段得到最终的预测结果。在UCF101以及Hmdb51数据集上实验准确率分别达到96.9%和76.2%,结果表明,非局部操作与scSE模块结合可以增强特征时空上以及通道间的信息提高准确率,验证了SC_NLResNet网络的有效性。

融合卷积注意力和Transformer架构的行人重识别方法

行人重识别技术是智能安防系统中的重要方法之一,为构建一个适用各种复杂场景的行人重识别模型,基于现有的卷积神经网络和Transformer模型,提出一种融合卷积注意力和Transformer(FCAT)架构的方法,以增强Transformer对局部细节信息的关注。所提方法主要将卷积空间注意力和通道注意力嵌入Transformer架构中,分别加强对图像中重要区域的关注和对重要通道特征的关注,以进一步提高Transformer架构对局部细节特征的提取能力。在3个公开行人重识别数据集上的对比消融实验证明,所提方法在非遮挡数据集上取得了与现有方法相当的结果,在遮挡数据集上的性能得到显著提升。所提方法更加轻量化,在不增加额外计算量和模型参数的情况下,推理速度得到了提升。

基于ConvLSTM-CNN预测太平洋长鳍金枪鱼时空分布趋势

海洋渔场的变动由空间与环境因子共同驱动,渔场时空演变信息的精准预测是海洋捕捞的关键。本研究利用1995-2018年太平洋海域长鳍金枪鱼(Thunnus alalunga)的渔业生产统计数据,结合同期海洋环境数据包括海表面温度(Sea Surface Temperature,SST)、海表面盐度(Sea Surface Salinity,SSS)、初级生产力(Primary Productivity,PP)和溶解氧浓度(Dissolved Oxygen Concentration,DO),提出了一种融合卷积长短期记忆网络(Convolutional Long Short-Term Memory Networks,ConvLSTM)和卷积神经网络(Convolutional Neural Networks,CNN)的渔场时空分布预测模型。该模型引入特征提取模块,对时空因子进行编码,提取时空特征信息,同时采用CNN提取海洋环境变量的抽象特征,采用ConvLSTM提取渔业数据的高层时空关联信息,最后融合多种特征对渔场时空演变趋势进行预测。结果表明,模型的均方根误差为0.1036,较随机森林、BP神经网络和长短期记忆网络(Long Short Term Memory,LSTM)等传统渔场预报模型的预测误差降低15%~40%,预测的高产渔区与实际作业的高渔获量区匹配度为89%。该研究构建的渔场时空预测模型能够准确地预测出太平洋长鳍金枪鱼的时空分布,为太平洋长鳍金枪鱼的延绳钓渔业提供科学参考依据。

适用于多跳TSN网络节点的帧抢占分片与重组方法

当前标准时间敏感网络中,帧抢占分片与重组机制仅适用于网络节点间。而可抢占帧在多跳时间敏感网络传输过程中,可能会多次出现帧抢占分片与重组,导致网络节点处理开销增加及转发效率降低的问题。设计了一种适用于多跳时间敏感网络节点传输的帧抢占分片与重组方法,通过择机分片、增加以太网地址、一次分片与重组、分片大小适当、保护带长度折中等措施,可减少多次帧抢占分片与重组带来的开销,提高网络节点转发效率,降低时间敏感流的时延和抖动,保障其传输时延的确定性。

融合GhostNet的YOLOv4轻量化网络设计与实现

由于存储资源和功耗等限制,在嵌入式设备上部署基于深度学习的目标检测算法存在困难,且效果不佳.基于YOLOv4算法,提出了一种改进的YOLOv4-Light轻量化网络模型,采用GhostNet网络结构替换CSPDarknet53作为主干提取网络,引入CBAM注意力机制关注通道和空间两个维度的特征信息,并利用感知量化方法对权重进行INT8量化处理,在保证精度的情况下降低网络模型规模和参数量.在PC端和NVIDIA Jetson Xavier NX上选用VisDrone无人机数据集分别对网络模型进行测试,结果表明YOLOv4-GhostNet-CBAM模型的尺寸是160M,比YOLOv4降低了34.43%;检测速率最高可达到34.6FPS,比YOLOv4提高了56.6%.YOLO-Light模型的尺寸是40.2M,比YOLOv4降低了83.5%;检测速率最高可达到78.6FPS,为YOLOv4的3.6倍,且交并比为0.5时的平均精度均值(mAP50)与YOLOv4相比仅下降了3%.YOLO-Light模型相较于原模型优势明显,能够在低功耗的嵌入式设备上完成实时目标检测.