08 图嵌入
文章目录
图嵌入
目标:想要嵌入一个子图或整个图,图形嵌入: Z G Z_G ZG
任务:
分类有毒分子和无毒分子,
识别异常图等
-方法1
简单的想法1:
在(子)图上运行标准的图形嵌入技术,然后只计算(子)图中的节点嵌入的总和(或平均值)。
2016年,杜维诺等人根据分子的图结构对分子进行分类。
-方法2
想法2:引入“虚拟节点”来表示(子)图,并运行标准图嵌入技术
由Li等人,2016年提出,作为子图嵌入的一般技术.
-方法3 Anonymous Walks Embedding
States in anonymous walks correspond to the index of the first time we visited the node in a random walk
Anonymous Walk Embeddings, ICML 2018 https://arxiv.org/pdf/1805.11921.pdf
忽略所访问节点的身份(因此是匿名)
匿名随机随机游走的序列数量
匿名随机游走的数量呈指数级增长:
There are 5 anon. walks w i w_i wi of length 3:
w 1 w_1 w1=111, w 2 w_2 w2=112, w 3 w_3 w3= 121, w 4 w_4 w4= 122, w 5 w_5 w5= 123
使用Anonymous Walks 嵌入图形
-Simulate anonymous walks w i w_i wi of l l l steps and record their counts
-Represent the graph as a probability distribution over these walks
我们可以进一步增强这个想法
New Idea: Learn Walk Embeddings
我们不是简单地表示每次行走的次数,而是学习嵌入
如何嵌入呢?(s.t.:使得,满足于)
Learn Walk Embeddings
总结
We discussed 3 ideas to graph embeddings
Approach 1: Embed nodes and sum/avg them
Approach 2: Create super-node that spans the (sub) graph and then embed that node
Approach 3: Anonymous Walk Embeddings
Idea 1: Sample the anon. walks and represent the graph as fraction of times each anon walk occurs
Idea 2: Embed anonymous walks, concatenate their embeddings to get a graph embedding