Graph-Augmentation-Free Self-Supervised Learning for Social Recommendation

Social recommendation systems can improve quality in cases of sparse user–item interaction data, which has attracted the industry’s attention. In reality, social mostly mine real user preferences from networks. However, trust relationships networks are complex and it is difficult to extract valuable preference information, worsens performance. To address this problem, paper proposes a algorithm based on multi-graph contrastive learning. ensure reliability preferences, builds multiple enhanced relationship views user’s network encodes multi-view high-order learning node representations using graph hypergraph convolutional Considering effect long-tail phenomenon, graph-augmentation-free self-supervised used as an auxiliary task contrastively enhance by adding uniform noise each layer encoder embeddings. Three open datasets were evaluate algorithm, was compared well-known systems. The experimental studies demonstrated superiority algorithm.