This report synthesises findings from 12 peer-reviewed papers addressing the following research question: What is the relative performance of LightGCL versus SimGCL in terms of recommendation accuracy (e.g., Recall@K, NDCG@K) when trained on large-scale sparse interaction graphs with varying levels of. Graph neural network (GNN) is a powerful learning approach for graph-based recommender systems. Recently, GNNs integrated with contrastive learning have shown superior performance in recommendation with their data augmentation schemes, aiming at dealing with highly sparse data. 9 claims were extracted from source literature; 9 were independently verified against retrieved documents. An automated multi-reviewer quality assessment produced a score of 8.8/10. This report is a machine-generated literature synthesis and does not constitute original research.Research goal: What is the relative performance of LightGCL versus SimGCL in terms of recommendation accuracy (e.g., Recall@K, NDCG@K) when trained on large-scale sparse interaction graphs with varying levels of sparsity?Autonomous literature synthesis. Automated review score: 8.8/10. Full text and citation available at Assignee Research.