Contents

PyTorch Geometric (PyG) is a geometric deep learning extension library for PyTorch.

torch_geometric.data

共以下十个类:

  • 单(个/批)图数据:
    • Data: A plain old python object modeling a single graph with various (optional) attributes
    • Batch: A plain old python object modeling a batch of graphs as one big (dicconnected) graph.
      • With torch_geometric.data.Data being the base class, all its methods can also be used here.
      • In addition, single graphs can be reconstructed via the assignment vector batch, which maps each node to its respective graph identifier.
  • 为数据集创建提供的两个抽象类( 官方教程 ):
    • Dataset: Dataset base class for creating graph datasets.
    • InMemoryDataset: … fit completely into memory. ( torch_geometric.datasets 中的现有数据集多从此类继承)

Contents

图表示学习

传统的Graph Mining是人工定义各种Metric,如出入度。

图表示学习通过学习映射,来获得节点在低维向量空间的表示。演变过程:

  • 图结构信息:邻居节点
  • 基于随机游走的邻居序列
    • 转化为Word2Vec:每个随机游走路径被当作Word2Vec中的一个句子,每个节点则是Word2Vec中的一个词。如DeepWalk,Node2Vec,Metapath2Vec(异构图)
    • DeepWalk第一次将深度学习中的技术引入到图表示学习领域
  • 基于局部子图的邻居:图神经网络
    • 通过用谱图理论空间局部性重新定义图卷积

图神经网络

图卷积神经网络GCN

卷积的本质:加权求和,只是组合了加和乘法一种复杂一点的运算。

Contents

DAPASA: Detecting Android Piggybacked Apps through Sensitive Subgraph Analysis

Published in: IEEE Transactions on Information Forensics and Security ( Volume: 12 , Issue: 8 , Aug. 2017 )