Double-Scale Self-Supervised Hypergraph Learning for Group Recommendation 论文笔记
Double-Scale Self-Supervised Hypergraph Learning for Group Recommendation
topic:
(1)定义了用户层、组层超图(解决现有方法未考虑到的复杂交互)
(2)双粒度自监督体系(解决稀疏问题)
提出模型:
模型三部分:
Hierarchical hypergraph,通过将信息从用户级传递到组级,来描述组内和组外的用户交互
Double-scale self-supervised learning, 包含粗粒度和细粒度的节点dropout策略,以细化用户和群组表示,缓解数据稀疏问题;
Model Optimization, 统一了群体推荐和自监督学习的目标,增强了两项任务。
1、Hierarchical Hypergraph Convolution
1.1 User-Level hypergraph
User representation learning
$$pl$$是指第l层超图卷积网络的user embedding
$$Dul
Double-Scale Self-Supervised Hypergraph Learning for Group Recommendation 论文笔记最先出现在Python成神之路。
共有 0 条评论