Abstract:Social relations are leveraged to tackle the sparsity issue of user-item interaction data in recommendation under the assumption of social homophily. However, social recommendation paradigms predominantly focus on homophily based on user preferences. While social information can enhance recommendations, its alignment with user preferences is not guaranteed, thereby posing the risk of introducing informational redundancy. We empirically discover that social graphs in real recommendation data exhibit low preference-aware homophily, which limits the effect of social recommendation models. To comprehensively extract preference-aware homophily information latent in the social graph, we propose Social Heterophily-alleviating Rewiring (SHaRe), a data-centric framework for enhancing existing graph-based social recommendation models. We adopt Graph Rewiring technique to capture and add highly homophilic social relations, and cut low homophilic (or heterophilic) relations. To better refine the user representations from reliable social relations, we integrate a contrastive learning method into the training of SHaRe, aiming to calibrate the user representations for enhancing the result of Graph Rewiring. Experiments on real-world datasets show that the proposed framework not only exhibits enhanced performances across varying homophily ratios but also improves the performance of existing state-of-the-art (SOTA) social recommendation models.

What problem does this paper attempt to address?

### Problems the Paper Aims to Solve This paper aims to address the issue of low homophily in social recommendation systems. Specifically: 1. **Research Background**: - Social relationships are used to address the sparsity problem in user-item interaction data, with the underlying assumption of social homophily (i.e., users tend to connect with people who have similar interests). - However, existing social recommendation methods mainly focus on homophily based on user preferences, while in reality, the social graph does not always exhibit highly homophilous connections. 2. **Core Issues**: - Existing social recommendation models directly integrate the original social graph into the system or construct a hypergraph based on the original social graph. However, these methods fail to distinguish between reliable and unreliable connections in the social graph, thereby affecting the recommendation performance. - Experiments have found that social graphs in real-world datasets exhibit a low homophily ratio, meaning that user connections in the social graph do not well reflect the similarity in user preferences. 3. **Solution**: - A framework named Social Heterophily-alleviating Rewiring (SHaRe) is proposed, which uses graph rewiring techniques to increase high-homophily social relationships and cut off low-homophily relationships. - To further optimize user representations, a contrastive learning method is introduced to calibrate user representations, thereby enhancing the effect of graph rewiring. 4. **Objectives**: - To optimize the use of reliable social relationships in the social graph and mitigate the impact of unreliable relationships to improve recommendation performance. - Experiments demonstrate that this framework can significantly enhance the performance of existing social recommendation models under different homophily ratios. Through this framework, the paper aims to improve the accuracy and personalization of social recommendation systems.

Challenging Low Homophily in Social Recommendation

A Novel Social Recommendation Method Fusing User’s Social Status and Homophily Based on Matrix Factorization Techniques

Disentangled Modeling of Social Homophily and Influence for Social Recommendation

Inhomogeneous Social Recommendation with Hypergraph Convolutional Networks

Social Recommender Systems Based on Coupling Network Structure Analysis

Hybrid Recommendation Model Based on Social Media and Social Networks

Dual Homogeneity Hypergraph Motifs with Cross-view Contrastive Learning for Multiple Social Recommendations

Heterogeneous Information Diffusion Model for Social Recommendation.

Multi-Graph Heterogeneous Interaction Fusion for Social Recommendation

Hete-CF: Social-Based Collaborative Filtering Recommendation using Heterogeneous Relations

Hypergraph-Enhanced Self-Supervised Robust Graph Learning for Social Recommendation.

Disentangled Hypergraph Collaborative Filtering for Social Recommendation

Interactive Social Recommendation

Graph Heterogeneous Multi-Relational Recommendation

SSRES: A Student Academic Paper Social Recommendation Model Based on a Heterogeneous Graph Approach

Social recommendation across multiple relational domains.

Heterogeneous Meta-Path Graph Learning for Higher-order Social Recommendation

Modelling High-Order Social Relations for Item Recommendation (extended Abstract)

Personalized Recommender Systems Based on Social Relationships and Historical Behaviors

Social Recommendation with Strong and Weak Ties

An enhanced social matrix factorization model for recommendation based on social networks using social interaction factors