Tripartite Heterogeneous Graph Propagation for Large-scale Social Recommendation
This work addresses social recommendation problems for users and platforms by improving recommendation accuracy, but it is incremental as it builds on existing GNN methods with specific modifications.
The paper tackled the challenges of complex noisy connections, high heterogeneity, and oversmoothing in social recommendation using Graph Neural Networks by proposing Heterogeneous Graph Propagation (HGP), which uses a tripartite graph and personalized PageRank to achieve improved performance, with experimental results showing it outperforms baselines on a large-scale dataset of 1,645,279 nodes and 4,711,208 edges in terms of AUC and F1-score metrics.
Graph Neural Networks (GNNs) have been emerging as a promising method for relational representation including recommender systems. However, various challenging issues of social graphs hinder the practical usage of GNNs for social recommendation, such as their complex noisy connections and high heterogeneity. The oversmoothing of GNNs is an obstacle of GNN-based social recommendation as well. Here we propose a new graph embedding method Heterogeneous Graph Propagation (HGP) to tackle these issues. HGP uses a group-user-item tripartite graph as input to reduce the number of edges and the complexity of paths in a social graph. To solve the oversmoothing issue, HGP embeds nodes under a personalized PageRank based propagation scheme, separately for group-user graph and user-item graph. Node embeddings from each graph are integrated using an attention mechanism. We evaluate our HGP on a large-scale real-world dataset consisting of 1,645,279 nodes and 4,711,208 edges. The experimental results show that HGP outperforms several baselines in terms of AUC and F1-score metrics.