LightGCN
Xiangnan HeUniversity of Science and Technology of China, HeFei, ChinaKuan DengUniversity of Science and Technology of China, HeFei, ChinaXiang WangNational University of Singapore, Singapore, SingaporeYan LiBeijing Kuaishou Technology Co., Ltd., BeiJing, ChinaYongdong ZhangUniversity of Science and Technology of China, HeFei, ChinaMeng WangHefei University of Technology, Hefei, China
2020en
ABI
Abstract
Graph Convolution Network (GCN) has become new state-of-the-art for collaborative filtering. Nevertheless, the reasons of its effectiveness for recommendation are not well understood. Existing work that adapts GCN to recommendation lacks thorough ablation analyses on GCN, which is originally designed for graph classification tasks and equipped with many neural network operations. However, we empirically find that the two most common designs in GCNs -- feature transformation and nonlinear activation -- contribute little to the performance of collaborative filtering. Even worse, including them adds to the difficulty of training and degrades recommendation performance.
Identifiers
Citations and references
Cited by 20 references