Dynamic Multi-Graph Spatio-Temporal Graph Traffic Flow Prediction in Bangkok:An Application of a Continuous Convolutional Neural Network

The ability to accurately predict urban traffic flows is crucial for optimising city operations.Consequently,various methods for forecasting urban traffic have been developed,focusing on analysing historical data to understand complex mobility patterns.Deep learning techniques,such as graph neural networks(GNNs),are popular for their ability to capture spatio-temporal dependencies.However,these models often become overly complex due to the large number of hyper-parameters involved.In this study,we introduce Dynamic Multi-Graph Spatial-Temporal Graph Neural Ordinary Differential Equation Networks(DMST-GNODE),a framework based on ordinary differential equations(ODEs)that autonomously discovers effective spatial-temporal graph neural network(STGNN)architectures for traffic prediction tasks.The comparative analysis of DMST-GNODE and baseline models indicates that DMST-GNODE model demonstrates superior performance across multiple datasets,consistently achieving the lowest Root Mean Square Error(RMSE)and Mean Absolute Error(MAE)values,alongside the highest accuracy.On the BKK(Bangkok)dataset,it outperformed other models with an RMSE of 3.3165 and an accuracy of 0.9367 for a 20-min interval,maintaining this trend across 40 and 60 min.Similarly,on the PeMS08 dataset,DMST-GNODE achieved the best performance with an RMSE of 19.4863 and an accuracy of 0.9377 at 20 min,demonstrating its effectiveness over longer periods.The Los_Loop dataset results further emphasise this model’s advantage,with an RMSE of 3.3422 and an accuracy of 0.7643 at 20 min,consistently maintaining superiority across all time intervals.These numerical highlights indicate that DMST-GNODE not only outperforms baseline models but also achieves higher accuracy and lower errors across different time intervals and datasets.

Dynamic adaptive spatio-temporal graph network for COVID-19 forecasting

Appropriately characterising the mixed space-time relations of the contagion process caused by hybrid space and time factors remains the primary challenge in COVID-19 forecasting.However,in previous deep learning models for epidemic forecasting,spatial and temporal variations are captured separately.A unified model is developed to cover all spatio-temporal relations.However,this measure is insufficient for modelling the complex spatio-temporal relations of infectious disease transmission.A dynamic adaptive spatio-temporal graph network(DASTGN)is proposed based on attention mechanisms to improve prediction accuracy.In DASTGN,complex spatio-temporal relations are depicted by adaptively fusing the mixed space-time effects and dynamic space-time dependency structure.This dual-scale model considers the time-specific,space-specific,and direct effects of the propagation process at the fine-grained level.Furthermore,the model characterises impacts from various space-time neighbour blocks under time-varying interventions at the coarse-grained level.The performance comparisons on the three COVID-19 datasets reveal that DASTGN achieves state-of-the-art results with a maximum improvement of 17.092%in the root mean-square error and 11.563%in the mean absolute error.Experimental results indicate that the mechanisms of designing DASTGN can effectively detect some spreading characteristics of COVID-19.The spatio-temporal weight matrices learned in each proposed module reveal diffusion patterns in various scenarios.In conclusion,DASTGN has successfully captured the dynamic spatio-temporal variations of COVID-19,and considering multiple dynamic space-time relationships is essential in epidemic forecasting.

An Intelligent Framework for Resilience Recovery of FANETs with Spatio-Temporal Aggregation and Multi-Head Attention Mechanism

Due to the time-varying topology and possible disturbances in a conflict environment,it is still challenging to maintain the mission performance of flying Ad hoc networks(FANET),which limits the application of Unmanned Aerial Vehicle(UAV)swarms in harsh environments.This paper proposes an intelligent framework to quickly recover the cooperative coveragemission by aggregating the historical spatio-temporal network with the attention mechanism.The mission resilience metric is introduced in conjunction with connectivity and coverage status information to simplify the optimization model.A spatio-temporal node pooling method is proposed to ensure all node location features can be updated after destruction by capturing the temporal network structure.Combined with the corresponding Laplacian matrix as the hyperparameter,a recovery algorithm based on the multi-head attention graph network is designed to achieve rapid recovery.Simulation results showed that the proposed framework can facilitate rapid recovery of the connectivity and coverage more effectively compared to the existing studies.The results demonstrate that the average connectivity and coverage results is improved by 17.92%and 16.96%,respectively compared with the state-of-the-art model.Furthermore,by the ablation study,the contributions of each different improvement are compared.The proposed model can be used to support resilient network design for real-time mission execution.

Multi-Scale Location Attention Model for Spatio-Temporal Prediction of Disease Incidence

Due to the increasingly severe challenges brought by various epidemic diseases,people urgently need intelligent outbreak trend prediction.Predicting disease onset is very important to assist decision-making.Most of the exist-ing work fails to make full use of the temporal and spatial characteristics of epidemics,and also relies on multi-variate data for prediction.In this paper,we propose a Multi-Scale Location Attention Graph Neural Networks(MSLAGNN)based on a large number of Centers for Disease Control and Prevention(CDC)patient electronic medical records research sequence source data sets.In order to understand the geography and timeliness of infec-tious diseases,specific neural networks are used to extract the geography and timeliness of infectious diseases.In the model framework,the features of different periods are extracted by a multi-scale convolution module.At the same time,the propagation effects between regions are simulated by graph convolution and attention mechan-isms.We compare the proposed method with the most advanced statistical methods and deep learning models.Meanwhile,we conduct comparative experiments on data sets with different time lengths to observe the predic-tion performance of the model in the face of different degrees of data collection.We conduct extensive experi-ments on real-world epidemic-related data sets.The method has strong prediction performance and can be readily used for epidemic prediction.

RECONSTRUCTION OF ONE DIMENSIONAL MULTI-LAYERED MEDIA BY USING A TIME DOMAIN SIGNAL FLOW GRAPH TECHNIQUE

A novel inverse scattering method to reconstruct the permittivity profile of one-dimensional multi-layered media is proposed in this paper.Based on the equivalent network ofthe medium,a concept of time domain signal flow graph and its basic principles are introduced,from which the reflection coefficient of the medium in time domain can be shown to be a series ofDirac δ-functions(pulse responses).In terms of the pulse responses,we will reconstruct both thepermittivity and the thickness of each layer will accurately be reconstructed.Numerical examplesverify the applicability of this

A network security situation awareness method based on layered attack graph

The real-time of network security situation awareness(NSSA)is always affected by the state explosion problem.To solve this problem,a new NSSA method based on layered attack graph(LAG)is proposed.Firstly,network is divided into several logical subnets by community discovery algorithm.The logical subnets and connections between them constitute the logical network.Then,based on the original and logical networks,the selection of attack path is optimized according to the monotonic principle of attack behavior.The proposed method can sharply reduce the attack path scale and hence tackle the state explosion problem in NSSA.The experiments results show that the generation of attack paths by this method consumes 0.029 s while the counterparts by other methods are more than 56 s.Meanwhile,this method can give the same security strategy with other methods.

Extracting multiple layers from data having graph structures

Much data such as geometric image data and drawings have graph structures. Such data are called graph structured data. In order to manage efficiently such graph structured data, we need to analyze and abstract graph structures of such data. The purpose of this paper is to find knowledge representations which indicate plural abstractions of graph structured data. Firstly, we introduce a term graph as a graph pattern having structural variables, and a substitution over term graphs which is graph rewriting system. Next, for a graph G, we define a multiple layer ( g,(θ 1,…,θ k )) of G as a pair of a term graph g and a list of k substitutions θ 1,…,θ k such that G can be obtained from g by applying substitutions θ 1,…,θ k to g. In the same way, for a set S of graphs, we also define a multiple layer for S as a pair ( D,Θ ) of a set D of term graphs and a list Θ of substitutions. Secondly, for a graph G and a set S of graphs, we present effective algorithms for extracting minimal multiple layers of G and S which give us stratifying abstractions of G and S, respectively. Finally, we report experimental results obtained by applying our algorithms to both artificial data and drawings of power plants which are real world data.

融合层注意力机制的多视角图对比学习推荐方法

图对比学习因其可有效缓解数据稀疏问题被广泛应用在推荐系统中.然而,目前大多数基于图对比学习的推荐算法均采用单一视角进行学习,这极大地限制了模型的泛化能力,且图卷积网络本身存在的过度平滑问题也影响着模型的稳定性.基于此,提出一种融合层注意力机制的多视角图对比学习推荐方法.一方面,该方法提出2种不同视角下的3种对比学习,在视图级视角下,通过对原始图添加随机噪声构建扰动增强视图,利用奇异值分解(singular value decomposition)重组构建SVD增强视图,对这2个增强视图进行视图级对比学习;在节点视角下,利用节点间的语义信息分别进行候选节点和候选结构邻居对比学习,并将3种对比学习辅助任务和推荐任务进行多任务学习优化,以提高节点嵌入的质量,从而提升模型的泛化能力.另一方面,在图卷积网络学习用户和项目的节点嵌入时,采用层注意力机制的方式聚合最终的节点嵌入,提高模型的高阶连通性,以缓解过度平滑问题.在4个公开数据集LastFM,Gowalla,Ifashion,Yelp上与10个经典模型进行对比,结果表明该方法在Recall,Precision,NDCG这3个指标上分别平均提升3.12%,3.22%,4.06%,这说明所提方法是有效的.

An attention graph stacked autoencoder for anomaly detection of electro-mechanical actuator using spatio-temporal multivariate signals

Health monitoring of electro-mechanical actuator(EMA)is critical to ensure the security of airplanes.It is difficult or even impossible to collect enough labeled failure or degradation data from actual EMA.The autoencoder based on reconstruction loss is a popular model that can carry out anomaly detection with only consideration of normal training data,while it fails to capture spatio-temporal information from multivariate time series signals of multiple monitoring sensors.To mine the spatio-temporal information from multivariate time series signals,this paper proposes an attention graph stacked autoencoder for EMA anomaly detection.Firstly,attention graph con-volution is introduced into autoencoder to convolve temporal information from neighbor features to current features based on different weight attentions.Secondly,stacked autoencoder is applied to mine spatial information from those new aggregated temporal features.Finally,based on the bench-mark reconstruction loss of normal training data,different health thresholds calculated by several statistic indicators can carry out anomaly detection for new testing data.In comparison with tra-ditional stacked autoencoder,the proposed model could obtain higher fault detection rate and lower false alarm rate in EMA anomaly detection experiment.

Domain Decomposition for Wavelet Single Layer on Geometries with Patches

We focus on the single layer formulation which provides an integral equation of the first kind that is very badly conditioned. The condition number of the unpreconditioned system increases exponentially with the multiscale levels. A remedy utilizing overlapping domain decompositions applied to the Boundary Element Method by means of wavelets is examined. The width of the overlapping of the subdomains plays an important role in the estimation of the eigenvalues as well as the condition number of the additive domain decomposition operator. We examine the convergence analysis of the domain decomposition method which depends on the wavelet levels and on the size of the subdomain overlaps. Our theoretical results related to the additive Schwarz method are corroborated by numerical outputs.

Distributed Contact Plan Design for Multi-Layer Satellite-Terrestrial Network

In multi-layer satellite-terrestrial network, Contact Graph Routing(CGR) uses the contact information among satellites to compute routes. However, due to the resource constraints in satellites, it is extravagant to configure lots of the potential contacts into contact plans. What's more, a huge contact plan makes the computing more complex, which further increases computing time. As a result, how to design an efficient contact plan becomes crucial for multi-layer satellite network, which usually has a large scaled topology. In this paper, we propose a distributed contact plan design scheme for multi-layer satellite network by dividing a large contact plan into several partial parts. Meanwhile, a duration based inter-layer contact selection algorithm is proposed to handle contacts disruption problem. The performance of the proposed design was evaluated on our Identifier/Locator split based satellite-terrestrial network testbed with 79 simulation nodes. Experiments showed that the proposed design is able to reduce the data delivery delay.

A Researchon the Problems Encountered in the Design of Hypertext Navigation Graph

Some problems encountered in developing navigational graph controlling program in ORACLE multimedia graphic development tool-Graphics such as multi-window creation,button simulation,computing descendant number and writing text etc.arc discussed,Since all kinds of algorithm related with the problems have been checked and proved to be correct,they have the feature of universal significance.

基于动态自适应图神经网络的电动汽车充电负荷预测

电动汽车充电站负荷波动的不确定性与长时间预测任务给提升充电负荷预测精度带来巨大的挑战。文中提出一种基于动态自适应图神经网络的电动汽车充电负荷预测算法。首先,构建了一个充电负荷信息时空关联特征提取层,将多头注意力机制与自适应相关图结合生成具有时空关联性的综合特征表达式,以捕获充电站负荷的波动性;然后,将提取的特征输入到时空卷积层,捕获时间和空间之间的耦合关系;最后,通过切比雪夫多项式图卷积与多尺度时间卷积提升模型耦合长时间序列之间的能力。以Palo Alto数据集为例,与现有方法相比,所提算法在4种波动情况下的平均预测误差大幅降低。

基于端到端深度神经网络和图搜索的OCT图像视网膜层边界分割方法

视网膜层边界的形态变化是眼部视网膜疾病出现的重要标志,光学相干断层扫描(optical coherence tomography,OCT)图像可以捕捉其细微变化,基于OCT图像的视网膜层边界分割能够辅助相关疾病的临床判断.在OCT图像中,由于视网膜层边界的形态变化多样,其中与边界相关的关键信息如上下文信息和显著性边界信息等对层边界的判断和分割至关重要.然而已有分割方法缺乏对以上信息的考虑,导致边界不完整和不连续.针对以上问题,提出一种“由粗到细”的基于端到端深度神经网络和图搜索(graph search,GS)的OCT图像视网膜层边界分割方法,避免了非端到端方法中普遍存在的“断层”现象.在粗分割阶段,提出一种端到端的深度神经网络—注意力全局残差网络(attention global residual network,AGR-Net),以更充分和有效的方式提取上述关键信息.具体地,首先设计一个全局特征模块(global feature module,GFM),通过从图像的4个方向扫描以捕获OCT图像的全局上下文信息;其次,进一步将通道注意力模块(channel attention module,CAM)与全局特征模块串行组合并嵌入到主干网络中,以实现视网膜层及其边界的上下文信息的显著性建模,有效解决OCT图像中由于视网膜层形变和信息提取不充分所导致的误分割问题.在细分割阶段,采用图搜索算法去除AGR-Net粗分割结果中的孤立区域或和孔洞等,保持边界的固定拓扑结构和连续平滑,以实现整体分割结果的进一步优化,为医学临床的诊断提供更完整的参考.最后,在两个公开数据集上从不同的角度对所提出的方法进行性能评估,并与最新方法进行比较.对比实验结果也表明所提方法在分割精度和稳定性方面均优于现有方法.

面向图谱频繁关系模式挖掘的异质图神经网络

鉴于目前挖掘算法难以对知识图谱建模等问题,提出一种描述和提取节点范围内结构的异质图神经网络模型,旨在挖掘其中的频繁关系模式以及各结构的分布。该模型将关系信息作为节点特征输入,利用自编码机制与多头注意力机制保留原始结构信息,同时引入特征结构平移层将相同结构映射到同一空间中,以获得频繁出现的结构。实验结果表明,该模型可以更快地挖掘图谱关系模式以及各结构在图中的分布;同时在验证特征表达能力的链接预测任务中有稳定表现,在关系类型较多的异质图中甚至优于部分联合学习模型。

基于图神经网络的多层银企网络融合研究

针对金融行业内潜在系统性风险难以精准识别问题,基于直接系统性风险传染渠道的借贷数据以及间接渠道的互联网文本信息,构建多层银企网络,并利用图卷积神经网络(GCN)设计多层银企网络融合模型,根据融合网络量化评估29家银行和75家房地产机构的不同渠道系统性风险传染过程。实验结果表明,在多层金融网络融合任务上,本文融合模型的准确率达到0.8559,优于对比模型。融合网络分析表明,多层网络共同冲击下的银企系统性风险传染能力明显大于单一或者2层网络的系统性风险,且基于间接渠道的企业间网络系统性风险更明显。金融审慎监管应该更多关注文本数据、深度学习等技术对于整合庞大金融资源的能力和有效提高风险监测预警的能力。

基于深度学习的工艺知识图谱构建及其应用

针对现有的零件工艺知识分散度高、结构性弱、复用性差等问题,提出了一种基于深度学习技术的工艺知识图谱构建和工艺重用方法。首先,分析了工艺知识结构,并建立了工艺知识图谱模式层;其次,搭建了深度学习知识抽取算法,并以工艺知识图谱模式层作为数据模式抽取了工艺知识,建立了工艺知识图谱的数据层;然后,基于图神经网络深度学习算法,搭建了工艺知识推理模型,将其作为工艺推荐基础;最后,搭建了零件工艺知识图谱可视化系统,并以行星架类零件为例,验证了工艺知识的检索和推荐功能。研究结果表明:该方法在工艺知识上的识别准确率达到了80%以上,工艺推荐准确率达到了70%以上,相比以往模型有所提高,证明了该方法在工艺知识图谱自动化构建和工艺重用上的有效性和可行性。

基于时空特征聚类和双层动态图卷积网络建模的短期净负荷预测

净负荷是实际负荷与光伏出力之差,针对净负荷中实际负荷强波动性与光伏出力强随机性相互耦合、表后光伏出力不可见等特点导致准确预测困难的问题,提出了一种基于时空特征聚类和双层动态图卷积网络建模的短期净负荷预测方法。首先,通过提取用户净负荷的日内时间特征、长期趋势特征和空间关联特征建立净负荷子集群聚类模型;其次,以子集群为图节点构建考虑“负荷-光伏”双维相关性的图结构,使其能够同时反映负荷和光伏出力特性;最后,引入净负荷总节点和动态邻接矩阵,构建通过长短期记忆神经网络连接的双层动态图卷积模型,得到净负荷预测结果。基于悉尼Ausgrid实际净负荷数据设计的消融实验结果表明,所提时空特征聚类方法和双层动态图结构分别使净负荷预测结果的均方根误差降低了13.44 kW和7.55 kW。未来将进一步拓展预测尺度,为电网保供决策提供更多信息支撑。

集改进图卷积和多层池化的点云分类模型

针对基于图卷积的点云分类模型在提取点云不同语义区域的特征信息以及高效利用聚合的高维特征方面存在的问题,本文提出了一种新的点云分类模型,该模型采用了动态自适应图卷积和多层池化相结合的方法。具体而言,本文采用了残差结构来构建更深层的卷积,以学习不同语义区域点对特征中不同层次的特征信息,从而生成动态自适应调整卷积核,针对不同的点对动态更新边的特征关系,从而提取更为精确的局部特征。同时,本文将聚合的高维特征输入到多层最大池化模块中,回收利用第一次最大池化后丢弃的特征信息进行多层最大池化,从而获取更为丰富的高维特征,提高分类模型的精度。实验结果表明,在ModelNet40数据集上,本文提出的分类模型的总体精度达到93.3%,平均精度为90.7%,明显优于目前主流的点云分类模型,并具有较强的鲁棒性。

基于三层数据治理的青年科技人才知识图谱构建与应用实践——以湖南省科技管理系统青年科技人才为例

[目的/意义]研究构建科技人才知识图谱,实现高潜质科技人才的智能化分析与识别,为科技人才评价、人才梯度培养等提供决策支撑。[方法/过程]本文从青年科技人才的特征出发,融合多源异构科技大数据,设计形成了科技人才知识图谱技术架构。从成长经历、科研环境、创新能力、科技领域等维度,构建了基于三层科技数据治理的青年科技人才知识图谱技术架构,建立了湖南青年人才数据资源池和画像系统。[结果/结论]本研究构建了涵盖湖南科技管理系统2万余条青年科技人才实体和40万条关系数据的湖南省青年科技人才知识图谱,有效支撑了湖南省青年科技人才管理决策。