Revisiting Graph based Collaborative Filtering: A Linear Residual Graph Convolutional Network Approach

By Lei Chen, Le Wu, Richang Hong, Kun Zhang, Meng Wang. AAAI Conference on Artificial Intelligence, 2020

654 citations recorded by the source corpus.

Graph Convolutional Networks~(GCNs) are state-of-the-art graph based representation learning models by iteratively stacking multiple layers of convolution aggregation operations and non-linear activation operations. Recently, in Collaborative Filtering~(CF) based Recommender Systems~(RS), by treating the user-item interaction behavior as a bipartite graph, some researchers model higher-layer collaborative signals with GCNs. These GCN based recommender models show superior performance compared to traditional works. However, these models suffer from training difficulty with non-linear activations for large user-item graphs. Besides, most GCN based models could not model deeper layers due to the over smoothing effect with the graph convolution operation. In this paper, we revisit GCN based CF models from two aspects. First, we empirically show that removing non-linearities would enhance recommendation performance, which is consistent with the theories in simple graph convolutional networks. Second, we propose a residual network structure that is specifically designed for CF with user-item interaction modeling, which alleviates the over smoothing problem in graph convolution aggregation operation with sparse user-item interaction data. The proposed model is a linear model and it is easy to train, scale to large datasets, and yield better efficiency and effectiveness on two real datasets. We publish the source code at https://github.com/newlei/LR-GCCF.

Subfields: Artificial Intelligence, Information Systems, Social Psychology

View Paper: https://www.semanticscholar.org/paper/c9e432023a240b22f385cf1cc76ab75fbe084ee7

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

1. Xiangnan He
2. Xiang Wang
3. Hongzhi Yin
4. Liqiang Nie
5. Chao Huang
6. Lianghao Xia
7. Junliang Yu
8. Meng Wang

In Information Systems

1. Xiangnan He
2. Xiang Wang
3. Hongzhi Yin
4. Chao Huang
5. Liqiang Nie
6. Meng Wang
7. Lianghao Xia
8. Junliang Yu

In Social Psychology

1. Hongzhi Yin

Citations in the corpus

• LightGCN: Simplifying and Powering Graph Convolution Network for Recommendation (2020, background): “Lastly, another work conducted in the same time [4] also finds that the nonlinearity is unnecessary in NGCF and develops linear GCN model for CF.”
• LightGCL: Simple Yet Effective Graph Contrastive Learning for Recommendation (2023, methodology): “GNN-based Collaborative Filtering: • GCCF (Chen et al., 2020c) strengthens the GNN-based collaborative filtering by implementing a residual network and reducing the non-linear transformation.”
• A Survey on Accuracy-Oriented Neural Recommendation: From Collaborative Filtering to Information-Rich Recommendation (2021, methodology): “Thus, a natural extension is to mine the [20], PMLAM [21], pQCF [22], FAWMF [23] User Interacted items+UID (Free Embed+Heuristic Agg) Item IID (Free Embed) SVD++ [24] User Interacted items (Free Embed+Heuristic Agg) Item IID (Free Embed) NAIS [25] User Interacted items+UID (Free Embed+Heuristic Agg)”
• Hypergraph Contrastive Collaborative Filtering (2022, methodology): “The effect of 𝜏 is further studied in the following subsection.”
• Self-Supervised Multi-Channel Hypergraph Convolutional Network for Social Recommendation (2021, background): “The difference is that we follow the suggestion in [6, 14] to remove the learnable matrix for linear transformation and the nonlinear activation function (e.”
• Interest-aware Message-Passing GCN for Recommendation (2021, background): “[3] have empirically demonstrated that the user/item embeddings become more similar when stacking more layers in NGCF due to the over-smoothing effect.”
• Graph Neural Networks in Recommender Systems: A Survey (2020, methodology): “NDCG NGCF [155], LR-GCCF [18], LightGCN [55], AGCN [173], HashGNN [135], DGCN-BinCF [142], DHCF [66], DGCF [156], NIA-GCN [132], Multi-GCCF [133]”
• UltraGCN: Ultra Simplification of Graph Convolutional Networks for Recommendation (2021, methodology): “Despite the wide success of GCN in graph learning, several recent studies [4, 10, 17, 29] found that simplifying GCN appropriately can further boost the performance on CF tasks.”
• DiffNet++: A Neural Influence and Interest Diffusion Network for Social Recommendation (2020, methodology): “could alleviate the over-smoothing issue [7].”
• Fine-Grained Similarity Measurement between Educational Videos and Exercises (2020, methodology): “The ranking performance is evaluated with the widely used metric NDCG@K [5, 35]:”