The Learning route of GNN

近期，随着WWW 2020会议的临近，有很多paper放出来，发现这次WWW 2020被图神经网络攻占，占比非常大，可见其火爆程度。

关于图神经网络（GNN）的必读文章

计算机视觉顶会CVPR 2020在不久前公布了论文接收列表。本届CVPR共收到了6656篇有效投稿，接收1470篇，其接受率在逐年下降，今年接受率仅为22%。整理了CVPR 2020 图神经网络(GNN)相关的比较有意思的值得阅读的，主要涉及视频文本检索、人体解析、图像描述生成、人脸重构、Human-Object Interaction。

概述 LADIES（Layer-Dependent Importance Sampling for Training Deep and Large Graph Convolutional Networks）是一种用于训练 GNN 的新型采样算法。 它考虑先前采样的节点来计算依赖于层的采样概率。 LADIES基于上层中的采样节点，选择其邻居节点，据此计算重要性概率，并在其中采样固定数量的节点。 我们从理论上和实验上证明，我们提出的采样算法在时间、内存和准确性方面优于以前的采样方法。 您可以查看我们的 NeurIPS 2019 论文以获取更多详细信息。 设置 此实现基于 Pytorch 我们假设您使用的是安装了 pip 的 Python 3。 要运行代码，您需要以下依赖项： 我们将三个小的基准节点分类数据集 cora、pubmed 和 citeseer 上传到 /data 中以供使用。 您

A PyTorch implementation of "Cluster-GCN: An Efficient Algorithm for Training Deep and Large Graph Convolutional Networks" (KDD 2019).

Morgan.Claypool.Action.Programming.Languages.-.Synthesis.Lectures.on.Artificial.Intelligence.and.Machine.Learning.pdf

Graphs are useful data structures in complex real-life applications such as modeling physical systems, learning molecular fingerprints, controlling traffic networks, and recommending friends in social networks. However, these tasks require dealing with non-Euclidean graph data that contains rich relational information between elements and cannot be well handled by traditional deep learning models (e.g., convolutional neural networks (CNNs) or recurrent neural networks (RNNs)). Nodes in graphs usually contain useful feature information that cannot be well addressed in most unsupervised representation learning methods (e.g., network embedding methods). Graph neural networks (GNNs) are proposed to combine the feature information and the graph structure to learn better representations on graphs via feature propagation and aggregation. Due to its convincing performance and high interpretability, GNN has recently become a widely applied graph analysis tool.