graph embedding相关论文阅读与实践

最新推荐文章于 2022-04-08 11:36:15 发布

小小码农JACK

最新推荐文章于 2022-04-08 11:36:15 发布

阅读量523

点赞数 2

分类专栏：机器学习算法

本文链接：https://blog.csdn.net/haoshan4783/article/details/90750429

版权

机器学习算法专栏收录该内容

38 篇文章 1 订阅

订阅专栏

论文及效果记录

graph embedding 综述
https://cloud.tencent.com/developer/article/1427335

Graph Embedding Techniques, Applications, and Performance: A Survey

将现有方法分为三大类：因式分解、随机游走、深度学习

Graph Factorization: Distributed Large-scale Natural Graph Factorization

ALS:Large-scale Parallel Collaborative Filtering for the Netflix Prize
https://www.jianshu.com/p/fc250e969dcc

HOPE:Asymmetric Transitivity Preserving Graph Embedding
两个版本，第一版本时间复杂度是0(n^3),速度太慢，第二版本自己实现，可8小时迭代一版embedding,加上pmi矩阵替换权重为1的邻接矩阵，效果还不错

SGNS:Neural Word Embedding as Implicit Matrix Factorization

Consider for example a pair of relatively frequent words (high P(w) and P©) that occur only once together. There is strong evidence that the words tend not to appear together, resulting in a negative PMI value, and hence a negative matrix entry. On the other hand, a pair of frequent words (same P (w) and P ©) that is never observed occurring together in the corpus, will receive a value of 0. So the PPMI is presented.

论文用svd分解PMI矩阵来替代word2vec
http://www.readventurer.com/?p=755