Abstract:It has been shown that the effectiveness of graph convolutional network (GCN) for recommendation is attributed to the spectral graph filtering. Most GCN-based methods consist of a graph filter or followed by a low-rank mapping optimized based on supervised training. However, we show two limitations suppressing the power of graph filtering: (1) Lack of generality. Due to the varied noise distribution, graph filters fail to denoise sparse data where noise is scattered across all frequencies, while supervised training results in worse performance on dense data where noise is concentrated in middle frequencies that can be removed by graph filters without training. (2) Lack of expressive power. We theoretically show that linear GCN (LGCN) that is effective on collaborative filtering (CF) cannot generate arbitrary embeddings, implying the possibility that optimal data representation might be unreachable. To tackle the first limitation, we show close relation between noise distribution and the sharpness of spectrum where a sharper spectral distribution is more desirable causing data noise to be separable from important features without training. Based on this observation, we propose a generalized graph normalization G^2N to adjust the sharpness of spectral distribution in order to redistribute data noise to assure that it can be removed by graph filtering without training. As for the second limitation, we propose an individualized graph filter (IGF) adapting to the different confidence levels of the user preference that interactions can reflect, which is proved to be able to generate arbitrary embeddings. By simplifying LGCN, we further propose a simplified graph filtering (SGFCF) which only requires the top-K singular values for recommendation. Finally, experimental results on four datasets with different density settings demonstrate the effectiveness and efficiency of our proposed methods.

Graph Convolutional Network for Recommendation with Low-pass Collaborative Filters.

Graph Learning in Low Dimensional Space for Graph Convolutional Networks

How Powerful is Graph Convolution for Recommendation?

SocialGCN: an Efficient Graph Convolutional Network Based Model for Social Recommendation

RGCF: Refined Graph Convolution Collaborative Filtering with concise and expressive embedding

Collaborative Filtering Recommendation Method Based on Graph Convolutional Neural Networks

HGCH: A Hyperbolic Graph Convolution Network Model for Heterogeneous Collaborative Graph Recommendation

How Powerful is Graph Filtering for Recommendation

Cluster-based Graph Collaborative Filtering

Hypergraph Contrastive Collaborative Filtering

LightGCN: Simplifying and Powering Graph Convolution Network for Recommendation

Fully Hyperbolic Graph Convolution Network for Recommendation

Bilateral Filtering Graph Convolutional Network for Multi-relational Social Recommendation in the Power-law Networks

A Lightweight Method of Knowledge Graph Convolution Network for Collaborative Filtering

PolyCF: Towards the Optimal Spectral Graph Filters for Collaborative Filtering

Multi-Behavior Enhanced Heterogeneous Graph Convolutional Networks Recommendation Algorithm based on Feature-Interaction

HN-GCCF: High-order neighbor-enhanced graph convolutional collaborative filtering

IntentGC: a Scalable Graph Convolution Framework Fusing Heterogeneous Information for Recommendation

Localized Graph Collaborative Filtering

AF-GCN: Attribute-Fusing Graph Convolution Network for Recommendation

FourierKAN-GCF: Fourier Kolmogorov-Arnold Network -- An Effective and Efficient Feature Transformation for Graph Collaborative Filtering