Network Configuration Entity Extraction Method Based on Transformer with Multi-Head Attention Mechanism

Nowadays,ensuring thequality of networkserviceshas become increasingly vital.Experts are turning toknowledge graph technology,with a significant emphasis on entity extraction in the identification of device configurations.This research paper presents a novel entity extraction method that leverages a combination of active learning and attention mechanisms.Initially,an improved active learning approach is employed to select the most valuable unlabeled samples,which are subsequently submitted for expert labeling.This approach successfully addresses the problems of isolated points and sample redundancy within the network configuration sample set.Then the labeled samples are utilized to train the model for network configuration entity extraction.Furthermore,the multi-head self-attention of the transformer model is enhanced by introducing the Adaptive Weighting method based on the Laplace mixture distribution.This enhancement enables the transformer model to dynamically adapt its focus to words in various positions,displaying exceptional adaptability to abnormal data and further elevating the accuracy of the proposed model.Through comparisons with Random Sampling(RANDOM),Maximum Normalized Log-Probability(MNLP),Least Confidence(LC),Token Entrop(TE),and Entropy Query by Bagging(EQB),the proposed method,Entropy Query by Bagging and Maximum Influence Active Learning(EQBMIAL),achieves comparable performance with only 40% of the samples on both datasets,while other algorithms require 50% of the samples.Furthermore,the entity extraction algorithm with the Adaptive Weighted Multi-head Attention mechanism(AW-MHA)is compared with BILSTM-CRF,Mutil_Attention-Bilstm-Crf,Deep_Neural_Model_NER and BERT_Transformer,achieving precision rates of 75.98% and 98.32% on the two datasets,respectively.Statistical tests demonstrate the statistical significance and effectiveness of the proposed algorithms in this paper.

Tracking direct and indirect impact on technology and policy of transformative research via ego citation network

Purpose:The disseminating of academic knowledge to nonacademic audiences partly relies on the transition of subsequent citing papers.This study aims to investigate direct and indirect impact on technology and policy originating from transformative research based on ego citation network.Design/methodology/approach:Key Nobel Prize-winning publications(NPs)in fields of gene engineering and astrophysics are regarded as a proxy for transformative research.In this contribution,we introduce a network-structural indicator of citing patents to measure technological impact of a target article and use policy citations as a preliminary tool for policy impact.Findings:The results show that the impact on technology and policy of NPs are higher than that of their subsequent citation generations in gene engineering but not in astrophysics.Research limitations:The selection of Nobel Prizes is not balanced and the database used in this study,Dimensions,suffers from incompleteness and inaccuracy of citation links.Practical implications:Our findings provide useful clues to better understand the characteristics of transformative research in technological and policy impact.Originality/value:This study proposes a new framework to explore the direct and indirect impact on technology and policy originating from transformative research.

基于CNN-Swin Transformer Network的LPI雷达信号识别

针对在低信噪比(SNR)条件下,低截获概率雷达信号调制方式识别准确率低的问题,提出一种基于Transformer和卷积神经网络(CNN)的雷达信号识别方法。首先,引入Swin Transformer模型并在模型前端设计CNN特征提取层构建了CNN+Swin Transformer网络(CSTN),然后利用时频分析获取雷达信号的时频特征,对图像进行预处理后输入CSTN模型进行训练,由网络的底部到顶部不断提取图像更丰富的语义信息,最后通过Softmax分类器对六类不同调制方式信号进行分类识别。仿真实验表明:在SNR为-18 dB时,该方法对六类典型雷达信号的平均识别率达到了94.26%,证明了所提方法的可行性。

Model Agnostic Meta-Learning(MAML)-Based Ensemble Model for Accurate Detection of Wheat Diseases Using Vision Transformer and Graph Neural Networks

Wheat is a critical crop,extensively consumed worldwide,and its production enhancement is essential to meet escalating demand.The presence of diseases like stem rust,leaf rust,yellow rust,and tan spot significantly diminishes wheat yield,making the early and precise identification of these diseases vital for effective disease management.With advancements in deep learning algorithms,researchers have proposed many methods for the automated detection of disease pathogens;however,accurately detectingmultiple disease pathogens simultaneously remains a challenge.This challenge arises due to the scarcity of RGB images for multiple diseases,class imbalance in existing public datasets,and the difficulty in extracting features that discriminate between multiple classes of disease pathogens.In this research,a novel method is proposed based on Transfer Generative Adversarial Networks for augmenting existing data,thereby overcoming the problems of class imbalance and data scarcity.This study proposes a customized architecture of Vision Transformers(ViT),where the feature vector is obtained by concatenating features extracted from the custom ViT and Graph Neural Networks.This paper also proposes a Model AgnosticMeta Learning(MAML)based ensemble classifier for accurate classification.The proposedmodel,validated on public datasets for wheat disease pathogen classification,achieved a test accuracy of 99.20%and an F1-score of 97.95%.Compared with existing state-of-the-art methods,this proposed model outperforms in terms of accuracy,F1-score,and the number of disease pathogens detection.In future,more diseases can be included for detection along with some other modalities like pests and weed.

Research on the longitudinal protection of a through-type cophase traction direct power supply system based on the empirical wavelet transform

This paper proposes a longitudinal protection scheme utilizing empirical wavelet transform(EWT)for a through-type cophase traction direct power supply system,where both sides of a traction network line exhibit a distinctive boundary structure.This approach capitalizes on the boundary’s capacity to attenuate the high-frequency component of fault signals,resulting in a variation in the high-frequency transient energy ratio when faults occur inside or outside the line.During internal line faults,the high-frequency transient energy at the checkpoints located at both ends surpasses that of its neighboring lines.Conversely,for faults external to the line,the energy is lower compared to adjacent lines.EWT is employed to decompose the collected fault current signals,allowing access to the high-frequency transient energy.The longitudinal protection for the traction network line is established based on disparities between both ends of the traction network line and the high-frequency transient energy on either side of the boundary.Moreover,simulation verification through experimental results demonstrates the effectiveness of the proposed protection scheme across various initial fault angles,distances to faults,and fault transition resistances.

Backflow Transformation for A=3 Nuclei with Artificial Neural Networks

A novel variational wave function defined as a Jastrow factor multiplying a backflow transformed Slater determinant was developed for A=3 nuclei.The Jastrow factor and backflow transformation were represented by artificial neural networks.With this newly developed wave function,variational Monte Carlo calculations were carried out for3H and3He nuclei starting from a nuclear Hamiltonian based on the leadingorder pionless effective field theory.The obtained ground-state energy and charge radii were successfully benchmarked against the results of the highly-accurate hypersphericalharmonics method.The backflow transformation plays a crucial role in improving the nodal surface of the Slater determinant and,thus,providing accurate ground-state energy.

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。

引入轻量级Transformer的无人机视觉跟踪

随着无人机在军事和民用领域的广泛运用,对于高精度、低功耗智能无人机跟踪系统的需求日益增加。针对目标跟踪算法在无人机跟踪场景下很难平衡跟踪精度和跟踪速度的问题,提出一种引入轻量级Transformer的孪生网络无人机目标跟踪算法SiamLT。使用Transformer对AlexNet网络进行改进,在增加最小计算量的情况下捕获全局特征信息。在目标模板与搜索区域匹配方面,联合Transformer和深度互相关运算提出一种二元相关模块,同时捕获目标模板与搜索区域之间的局部相关性和全局依赖关系。在分类回归网络中引入距离交并比,并采用多监督策略训练网络,以获取更准确的目标位置。在UAV123和UAV20L跟踪基准上的实验结果表明,SiamLT算法优于主流的目标跟踪算法,更有效地平衡了跟踪精度和跟踪速度。

多尺度卷积结合Transformer的抑郁脑电分类研究

在通过深度学习模型进行抑郁症类脑电信号分析时,针对单一尺度的卷积存在特征提取不充分的问题和卷积神经网络在感知脑电信号全局依赖性方面的局限性,分别设计了多尺度动态卷积网络模块和门控Transformer编码器模块,并与时间卷积网络相结合,提出了混合网络模型(MGTTCNet)进行抑郁症患者和健康对照组的脑电信号分类。该模型首先通过多尺度动态卷积从空间域和频率域捕捉脑电信号的多尺度时频信息。其次通过门控Transformer编码器学习脑电信号中的全局依赖关系,其利用多头注意力机制有效增强网络表达相关脑电信号特征的能力。之后利用时间卷积网络提取脑电信号可用的时间特征,最后将提取的抽象特征输入到分类模块进行分类。在公开数据集MODMA上用留出法和十折交叉验证法对提出模型进行实验验证,分别取得了约98.51%和98.53%的分类准确率,相较于基线单尺度模型EEGNet,分类准确率分别提升了约1.89%和1.93%,F1值分别提升了约2.05%和2.08%,kappa系数值分别提高了约0.038 1和0.038 5;同时消融实验验证了文中设计的各个模块的有效性。

CNN-Transformer结合对比学习的高光谱与LiDAR数据协同分类

针对高光谱图像(hyperspectral images,HSI)与LiDAR数据多模态分类任务中的跨模态信息表达和特征对齐等问题,提出一种基于对比学习CNN-Transformer高光谱和LiDAR数据协同分类网络(Contrastive Learning based CNNTransformer Network,CLCT-Net)。CLCT-Net通过由ConvNeXt V2 Block构成的共有特征提取模块,获得不同模态间的共性特征,解决异构传感器数据之间语义对齐的问题。构建了包含空间-通道分支和光谱上下文分支的双分支HSI编码器,以及结合频域自注意力机制的LiDAR编码器,以获取更丰富的特征表示。利用集成对比学习进行分类,进一步提升多模态数据协同分类的精度。在Houston 2013和Trento数据集上的实验结果表明,相较于其他高光谱图像和Li‐DAR数据分类模型,本文所提模型获得了更高的地物分类精度,分别达到了92.01%和98.90%,实现了跨模态数据特征的深度挖掘和协同提取。

基于多模态掩码Transformer网络的社会事件分类

多模态社会事件分类的关键是充分且准确地利用图像和文字2种模态的特征。然而,现有的大多数方法存在以下局限性:简单地将事件的图像特征和文本特征连接起来,不同模态之间存在不相关的上下文信息导致相互干扰。因此,仅仅考虑多模态数据模态间的关系是不够的,还要考虑模态之间不相关的上下文信息(即区域或单词)。为克服这些局限性,提出一种新颖的基于多模态掩码Transformer网络(MMTN)模型的社会事件分类方法。通过图-文编码网络来学习文本和图像的更好的表示。将获得的图像和文本表示输入多模态掩码Transformer网络来融合多模态信息,并通过计算多模态信息之间的相似性,对多模态信息的模态间的关系进行建模,掩盖模态之间的不相关上下文。在2个基准数据集上的大量实验表明:所提模型达到了最先进的性能。

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Transformer和CNN交错混合的肺结节分割网络

针对肺结节尺寸多样、形状异质化高等问题,提出基于Transformer和卷积神经网络(CNN)交错混合(IMTC)的肺结节分割网络,该网络是一个对称的层次连接网络,具有很强的多尺度特征提取能力。该网络通过集成2种方案分别解决肺结节多尺寸与形状异质化问题:(1)采用感知注意力模块(inception attention module,IAM),通过并联多个不同大小的卷积核来增加浅层网络的感受野组合,以此捕获更为丰富的浅层特征;(2)为获取更具表示能力的高级语义特征,利用由Transformer和CNN组成的基本骨干网络交错提取结节特征,使得全局特征与局部特征充分融合,从而提高结节特征表示的泛化能力和鲁棒性。实验结果表明:本文模型可以准确分割直径较小以及边缘复杂的肺结节,在LUNA16公开数据集上分割性能良好,Dice和IOU分别达到86.15%和76.10%。

基于生成对抗Transformer的电力负荷数据异常检测

电力负荷异常数据将给电力系统规划、负荷预测以及用能分析等带来较大的负面影响,因此亟须对负荷数据异常进行检测与识别。首先,针对电力负荷数据异常分类、原因及其特征开展分析。其次,改进传统Transformer编码器结构,采用多头注意力层代替掩码多头注意力层,同时移除前馈网络,以提高模型对负荷时序序列的全局注意力。基于生成对抗网络(generative adversarial networks,GAN)生成器与判别器的博弈结构,提出一种改进的GAN-Transformer模型,以更好地捕捉趋势性特征并加速模型收敛。然后,引入多阶段映射与训练方法,综合焦点分数打分机制,通过分阶段负荷序列重构帮助模型更好地提取负荷数据异常特征。最后,算例分析结果表明,GAN-Transformer模型在负荷数据异常检测精确率、召回率、F_(1)值以及训练时间方面均具有更优的性能,验证了所提方法的有效性和优越性。文中研究工作为基于深度学习进一步实现电力负荷数据异常分类与数据修复提供了有益参考。

ConvFormer:基于Transformer的视觉主干网络

针对主流Transformer网络仅对输入像素块做自注意力计算而忽略了不同像素块间的信息交互,以及输入尺度单一导致局部特征细节模糊的问题,本文提出一种基于Transformer并用于处理视觉任务的主干网络ConvFormer. ConvFormer通过所设计的多尺度混洗自注意力模块(Channel-Shuffle and Multi-Scale attention,CSMS)和动态相对位置编码模块(Dynamic Relative Position Coding,DRPC)来聚合多尺度像素块间的语义信息,并在前馈网络中引入深度卷积提高网络的局部建模能力.在公开数据集ImageNet-1K,COCO 2017和ADE20K上分别进行图像分类、目标检测和语义分割实验,ConvFormer-Tiny与不同视觉任务中同量级最优网络RetNetY-4G,Swin-Tiny和ResNet50对比,精度分别提高0.3%,1.4%和0.5%.

基于Transformer网络多模态融合的密集视频描述方法

针对目前的密集视频描述模型大多使用两阶段的方法存在效率较低、忽略音频及语义信息,描述结果不全面的问题。提出了一种基于Transformer网络多模态和语义信息融合的密集视频描述方法。提取自适应R(2+1)D网络提取视觉特征,设计了语义探测器生成语义信息,加入音频特征进行补充,建立了多尺度可变形注意力模块,应用并行的预测头,加快模型收敛速度,提高模型精度。实验结果表明:模型在2个基准数据集上性能均有很好的表现,评价指标BLEU4上达到了2.17。

基于卷积神经网络与Transformer的电能质量扰动分类方法

复杂电能质量扰动(power quality disturbances, PQD)的智能分类对于智能电网发展具有重要意义。扰动特征的提取与定位、模式识别与分类是电能质量扰动分类方法研究的难点。采用深度学习算法,将具有关注全局信息的Transformer与善于提取局部特征的卷积神经网络相融合,提出一种基于卷积神经网络(convolutional neural network, CNN)与Transformer的电能质量扰动分类方法,即CTranCBA。这种双深度学习模型分类方法主要是通过一维卷积神经网络提取电能质量扰动信号特征,利用Transformer自注意力机制引导模型关注序列中不同位置间的依赖关系,实现对扰动信号局部特征与全局特征的互补,克服了因感受野的限制而带来的识别不清、分类不准等问题。使用23种不同电能质量扰动信号,将CTranCBA与Deep-CNN、CNN-LSTM、CNN-CBAM方法进行比较。结果表明:该方法在分类准确率和抗噪性方面表现优异,可为电能质量扰动智能分类提供一种新的方法。

复谱映射下融合高效Transformer的语音增强方法

针对卷积神经网络(Convolutional Neural Network,CNN)过去在语音增强中表现优异但对全局特征捕获不足,以及Transformer近年展现出长序列间依赖优势但又存在局部细节特征丢失、参数量大等问题,该文为了充分利用CNN与Transformer的优势并弥补各自不足,提出了一种在复频谱映射下的新型卷积模块与高效Transformer融合的单通道语音增强网络。该网络由编码层、传输层与双分支解码层组成:在编解码部分设计了一种协作学习模块(Collaborative Learning Block,CLB)来监督交互信息,在减少参数量的同时提高主干网络对复特征的获取能力;传输层中则提出一种时频空间注意Transformer模块分别对语音子频带和全频带信息建模,充分利用声学特性来模拟局部频谱模式并捕获谐波间依赖关系。将该模块进一步与通道注意分支相结合,设计了一种可学习的双分支注意融合(Dual-branch Attention Fusion,DAF)机制,从空间-通道角度提取上下文特征以加强信息的多维度传输;最后,在此基础上搭建一种高斯加权渐进网络作为中间传输层,通过堆叠DAF模块进行加权求和后输出以充分利用深层特征,使得解码过程更具鲁棒性。分别在英文VoiceBank-DEMAND数据集、中文THCHS30语料库与115种环境噪声下进行消融以及综合对比实验,结果表明,该文方法仅以最小0.68×10^(6)的参数量,相比于大部分最新相关网络模型取得了更优的主、客观指标,具有较为突出的增强性能与泛化能力。

融合Transformer和交互注意力网络的方面级情感分类模型

现有的大多数研究者使用循环神经网络与注意力机制相结合的方法进行方面级情感分类任务。然而,循环神经网络不能并行计算,并且模型在训练过程中会出现截断的反向传播、梯度消失和梯度爆炸等问题,传统的注意力机制可能会给句子中重要情感词分配较低的注意力权重。针对上述问题,该文提出了一种融合Transformer和交互注意力网络的方面级情感分类模型。首先利用BERT(bidirectional encoder representation from Transformers)预训练模型来构造词嵌入向量,然后使用Transformer编码器对输入的句子进行并行编码,接着使用上下文动态掩码和上下文动态权重机制来关注与特定方面词有重要语义关系的局部上下文信息。最后在5个英文数据集和4个中文评论数据集上的实验结果表明,该文所提模型在准确率和F1上均表现最优。

基于IWOA-Transformer的磨煤机故障预警

提出了一种基于改进鲸鱼算法优化Transformer网络超参数(IWOA-Transformer)的故障预警方法。该方法利用非线性收敛系数和高斯变异对鲸鱼算法(WOA)进行改进,以提高WOA的收敛速度和避免其陷入局部最优;再采用改进鲸鱼算法(IWOA)优化Transformer的超参数,建立磨煤机故障预警模型;然后,通过预测值和实际值的相似度函数确定自适应阈值,结合专家系统判断故障类型并提出解决方案,实现磨煤机故障预警;最后,以某350 MW热电机组中速磨煤机为例进行故障预警试验。结果表明:所提IWOA-Transformer模型可显著提高预警速度和准确率,具有工程实用价值。