| Autor: |
Peng, Shaowen, Sugiyama, Kazunari, Mine, Tsunenori |
| Zdroj: |
ACM Transactions on Information Systems; May2024, Vol. 42 Issue 3, p1-26, 26p |
| Abstrakt: |
The article focuses on reducing redundancy in Graph Convolutional Networks (GCNs) for recommendation systems, addressing feature, and distribution redundancies. Topics include unveiling the inefficiencies of existing GCN-based methods, proposing a Simplified Graph Denoising Encoder (SGDE) to enhance model efficiency, and introducing a scalable contrastive learning framework for improved robustness and generalization. |
| Databáze: |
Complementary Index |
| Externí odkaz: |
|
