Graph-less Collaborative Filtering
arXiv:2303.08537 · doi:10.1145/3543507.3583196
Abstract
Graph neural networks (GNNs) have shown the power in representation learning over graph-structured user-item interaction data for collaborative filtering (CF) task. However, with their inherently recursive message propagation among neighboring nodes, existing GNN-based CF models may generate indistinguishable and inaccurate user (item) representations due to the over-smoothing and noise effect with low-pass Laplacian smoothing operators. In addition, the recursive information propagation with the stacked aggregators in the entire graph structures may result in poor scalability in practical applications. Motivated by these limitations, we propose a simple and effective collaborative filtering model (SimRec) that marries the power of knowledge distillation and contrastive learning. In SimRec, adaptive transferring knowledge is enabled between the teacher GNN model and a lightweight student network, to not only preserve the global collaborative signals, but also address the over-smoothing issue with representation recalibration. Empirical results on public datasets show that SimRec archives better efficiency while maintaining superior recommendation performance compared with various strong baselines. Our implementations are publicly available at: https://github.com/HKUDS/SimRec.
Accepted by ACM WWW 2023
References in corpus (15)
- Distilling the Knowledge in a Neural Network
- BPR: Bayesian Personalized Ranking from Implicit Feedback
- Simplifying Graph Convolutional Networks
- Graph Convolutional Matrix Completion
- Disentangled Graph Collaborative Filtering
- Improving Graph Collaborative Filtering with Neighborhood-enriched Contrastive Learning
- Towards Deeper Graph Neural Networks
- Hypergraph Contrastive Collaborative Filtering
- Heterogeneous Graph Contrastive Learning for Recommendation
- Contrastive Meta Learning with Behavior Multiplicity for Recommendation
- Self-Supervised Hypergraph Transformer for Recommender Systems
- LightGCL: Simple Yet Effective Graph Contrastive Learning for Recommendation
- Towards Deeper Graph Neural Networks with Differentiable Group Normalization
- DE-RRD: A Knowledge Distillation Framework for Recommender System
- STAR-GCN: Stacked and Reconstructed Graph Convolutional Networks for Recommender Systems