期刊文献+

基于稀疏注意力的自适应图池化方法 被引量:2

Adaptive graph pooling method based on sparse attention
下载PDF
导出
摘要 针对图神经网络池化中不能充分保留图的局部特征的问题,提出一种基于稀疏注意力自适应图池化方法。首先,运用稀疏注意力自适应地选择相似度高的节点形成集群;然后,运用局部聚合卷积,通过节点聚合形成集群表示,选取集群表示最大的topk个节点完成采样;最后,在池化时保留图的局部特征以降低信息损失,从而提高图分类的性能。实验结果表明,与传统图池化方法相比,提出方法的分类正确率有所提升。 An adaptive graph pooling method based on sparse attention is proposed to solve the problem that the graph neural network pooling can not fully preserve the local features of the graph.Firstly,sparse attention is used to adaptively select nodes with high similarity to form clusters.Then,local aggregation convolution is used to form cluster representation through node aggregation.The nodes mapping to the largest top-k cluster representation are selected to complete the sampling.Finally,the local features of the graph are retained to reduce information loss and improve the performance of graph classification.Experimental results show that the proposed method has a higher classification accuracy than the prevailing graph pooling methods.
作者 朱小草 郭春生 张宏宽 金昊炫 ZHU Xiaocao;GUO Chunsheng;ZHANG Hongkuan;JIN Haoxuan(School of Communication Engineering,Hangzhou Dianzi University,Hangzhou Zhejiang 310018,China;Soyea Technology Co.,LTD.,Hangzhou Zhejiang 310012,China;Zhejiang Academy of Agricultural Sciences,Hangzhou Zhejiang 310021,China)
出处 《杭州电子科技大学学报(自然科学版)》 2021年第5期32-38,共7页 Journal of Hangzhou Dianzi University:Natural Sciences
关键词 图池化 稀疏注意力 自适应处理 graph pooling sparse attention adaptive processing
  • 相关文献

同被引文献24

引证文献2

二级引证文献1

相关作者

内容加载中请稍等...

相关机构

内容加载中请稍等...

相关主题

内容加载中请稍等...

浏览历史

内容加载中请稍等...
;
使用帮助 返回顶部