首先是global graph上的表征,遵循标准的GNN信息传递机制,采用了加权汇聚邻域结点,作者提出了session-aware的注意力汇聚机制,会计算vi在global graph上的每个邻居结点vj和vi的亲和度值,亲和度值的计算过程和目标session序列表征以及表征对象vi都有关;针对session graph,作者区分了多种连边关系,入度边,出度边,自连接边,...
【SR-GNN | AAAI 2019】:Session-Based Recommendation with Graph Neural Networks:基于图神经网络的会话推荐 【GC-SAN | IJCAI 2019】:Graph Contextualized Self-Attention Network for Session-based Recommendation:基于图上下文自我注意力网络的会话推荐 【CA-TAN | ICDM 2020】:Cross-Session Aware Temporal Convol...
1 个内容 【代码解读】Global Context Enhanced Graph Neural Networks for Session-based Recommendation Starry Starry: 20220126 第15篇续 注:本文为2020 SIGIR论文Global Context Enhanced Graph Neural Networks for Session-base… 阅读全文 赞同 12 ...