Convolutional Neural Networks on Graphs with Fast Localized Spectral Filtering:谱域 Semi-Supervised Classification with Graph Convolutional Networks:谱域 空域论文# Neural Network for Graphs: A Contextual Constructive Approach:空域图卷积早期代表作品 Diffusion-Convolutional Neural Networks:空域 Learning Convolutiona...
(1)Recurrent Graph Neural Networks: GNN的先驱,其目的是学习具有循环神经结构的节点表示,RecGNN假设图中的一个节点不断地与它的邻居交换信息/消息,直到达到稳定的均衡。 (2)Convolutional Graph Neural Networks: ConvGNN将网格数据的卷积运算推广到Graph数据。主要思想:聚合节点v自身的特征x_v和其邻居的特征x_u来...
(1)Recurrent Graph Neural Networks:GNN的先驱,其目的是学习具有循环神经结构的节点表示,RecGNN假设图中的一个节点不断地与它的邻居交换信息/消息,直到达到稳定的均衡。 (2)Convolutional Graph Neural Networks:ConvGNN将网格数据的卷积运算推广到Graph数据。主要思想:聚合节点 自身的特征 和其邻居的特征 来生成节点 ...
Index Terms—Deep Learning, graph neural networks, graph convolutional networks, graph representation learning, graph autoencoder, network embedding I. INTRODUCTION(引言) 神经网络最新的成就推动了对模式识别和数据挖掘的研究。许多机器学习任务,例如目标检测[1],[2],机器翻译[3],[4]和语音识别[5],这些任务...
Deep Convolutional Networks on Graph-Structured Data:谱域 Convolutional Neural Networks on Graphs with Fast Localized Spectral Filtering:谱域 Semi-Supervised Classification with Graph Convolutional Networks:谱域 空域论文 Neural Network for Graphs: A Contextual Constructive Approach:空域图卷积早期代表作品 ...
基于空间的图卷积神经网络Spatial-based Graph Convolutional Networks 门控注意力网络(Gated Attention Network)(GANN) 图形注意力模型(Graph Attention Model)(GAM) 图自动编码器(Graph Autoencoders) Graph Autoencoder (GAE)和Adversarially Regularized Graph Autoencoder (ARGA) ...
递归图神经网络 (recurrent graph neural networks, RecGNNs) 卷积图神经网络 (convolutional graph neural networks, ConvGNN) 图自编码器 (graph autoencoders, GAE) 考虑时间因素的图神经网络 (spatial-temporal graph neural networks, ST-GNN) 讨论图神经网络在各个领域的应用 ...
DNN(ConvolutionalNeuralNetwork,卷积神经网络) Why CNN for Image The same patterns appear in different regions Subsampling the pixels will not change the objectGNN(GraphNeuralNetworks) 通过节点把图像变成图(数据结构 2019年Philip S. Yu团队的图神经网络综述 ...
neuralnetworksacrossvariousdomainsandsummarizetheopensourcecodes andbenchmarksoftheexistingalgorithmsondifferentlearningtasks.Finally,weproposepotentialresearchdirectionsinthis fast-growingfield. IndexTerms—DeepLearning,graphneuralnetworks,graphconvolutionalnetworks,graphrepresentationlearning,graph autoencoder,network...
图循环神经网络 (graph recurrent neural networks,Graph RNN),Graph RNN 通过在节点级别或图级别的状态进行建模,来获得图的递归和序列特征。 图卷积网络 (graph convolutional networks,GCN),GCN 定义了对不规则图结构的卷积和读出操作(readout operation),以获得常见的局部和全局结构特征。