Graph Contrastive Learning for Optimizing Sparse Data in Recommender Systems with LightGCL
This addresses data sparsity and noise in recommender systems, with improvements in fairness and bias resilience, though it appears incremental as an enhancement to existing graph neural network methods.
The paper tackles data sparsity and noise in recommender systems by implementing LightGCL, a graph contrastive learning model that uses SVD for graph augmentation, achieving significant gains over state-of-the-art models on benchmark datasets.
Graph Neural Networks (GNNs) are powerful tools for recommendation systems, but they often struggle under data sparsity and noise. To address these issues, we implemented LightGCL, a graph contrastive learning model that uses Singular Value Decomposition (SVD) for robust graph augmentation, preserving semantic integrity without relying on stochastic or heuristic perturbations. LightGCL enables structural refinement and captures global collaborative signals, achieving significant gains over state-of-the-art models across benchmark datasets. Our experiments also demonstrate improved fairness and resilience to popularity bias, making it well-suited for real-world recommender systems.