Abstract:Social recommendation has achieved great success in many domains including e-commerce and location-based social networks. Existing methods usually explore the user-item interactions or user-user connections to predict users’ preference behaviors. However, they usually learn both user and item representations in Euclidean space, which has large limitations for exploring the latent hierarchical property in the data. In this article, we study a novel problem of hyperbolic social recommendation, where we aim to learn the compact but strong representations for both users and items. Meanwhile, this work also addresses two critical domain-issues, which are under-explored. First, users often make trade-offs with multiple underlying aspect factors to make decisions during their interactions with items. Second, users generally build connections with others in terms of different aspects, which produces different influences with aspects in social network. To this end, we propose a novel graph neural network (GNN) framework with multiple aspect learning, namely, HyperSoRec. Specifically, we first embed all users, items, and aspects into hyperbolic space with superior representations to ensure their hierarchical properties. Then, we adapt a GNN with novel multi-aspect message-passing-receiving mechanism to capture different influences among users. Next, to characterize the multi-aspect interactions of users on items, we propose an adaptive hyperbolic metric learning method by introducing learnable interactive relations among different aspects. Finally, we utilize the hyperbolic translational distance to measure the plausibility in each user-item pair for recommendation. Experimental results on two public datasets clearly demonstrate that our HyperSoRec not only achieves significant improvement for recommendation performance but also shows better representation ability in hyperbolic space with strong robustness and reliability.

What problem does this paper attempt to address?

The main problem that this paper attempts to solve is that existing social recommendation systems usually use Euclidean space when modeling user and item representations, which fails to fully capture the potential hierarchical structure characteristics in the data. Specifically, both the user - item interaction graph and the user - user connection network exhibit a tree - like structure, and this structure is difficult to be effectively embedded in Euclidean space. Therefore, the paper proposes a new problem: **Hyperbolic Social Recommendation**, aiming to improve the recommendation performance by learning compact and powerful user and item representations. ### Specific Problem Description 1. **Hierarchical Structure Characteristics**: - The degree of each user or item node in the user - item interaction graph usually follows a power - law distribution, which can be traced back to the hierarchical structure. - The user - user connection network also shows an underlying tree - like structure, that is, the number of users connected to the central user grows exponentially. 2. **Multifaceted Influences**: - Users will weigh factors in multiple potential aspects when interacting with items. For example, when choosing a mobile phone, factors such as "price", "brand" and "appearance" may be considered. - When users establish connections with others in the social network, different influences will also be generated according to different aspects. For example, "User A" may trust "Friend B" more regarding brand suggestions, but rely more on "Friend C" regarding price opinions. ### Solution To solve the above problems, the paper proposes a novel graph neural network framework - **HyperSoRec**, which combines a multifaceted learning mechanism to perform user and item representation learning in hyperbolic space. The specific steps are as follows: 1. **Hyperbolic Space Embedding**: - Embed all users, items and aspects into hyperbolic space to ensure their hierarchical structure characteristics. 2. **Multifaceted Message Passing Mechanism**: - Adapt to the graph neural network and introduce a new multifaceted message passing and receiving mechanism to capture the influences of different aspects. 3. **Adaptive Hyperbolic Metric Learning**: - Propose an adaptive hyperbolic metric learning method, and characterize the multifaceted influences of users on items by introducing learnable interaction relationships. 4. **Translation Distance Calculation**: - Use hyperbolic translation distance to measure the rationality of each user - item pair. ### Experimental Verification The experimental results show that HyperSoRec not only significantly improves the recommendation performance, but also shows better representation ability, stronger robustness and reliability in hyperbolic space. In conclusion, by introducing hyperbolic space and multifaceted learning mechanisms, this paper solves the deficiencies of existing social recommendation systems in capturing hierarchical structure characteristics and multifaceted influences, thereby improving the performance of the recommendation system.

HyperSoRec: Exploiting Hyperbolic User and Item Representations with Multiple Aspects for Social-aware Recommendation

A Hyperbolic-to-Hyperbolic User Representation with Multi-aspect for Social Recommendation

Automatic Hypergraph Generation for Enhancing Recommendation with Sparse Optimization

Thinking inside The Box: Learning Hypercube Representations for Group Recommendation

Fully Hyperbolic Graph Convolution Network for Recommendation

HSR: Hyperbolic Social Recommender

Inhomogeneous Social Recommendation with Hypergraph Convolutional Networks

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

HSA: Hyperbolic Self-attention for Sequential Recommendation

HyperSegRec: enhanced hypergraph-based recommendation system with user segmentation and item similarity learning

A Recommendation Method Based on Multi-Source Heterogeneous Hypergraphs and Contrastive Learning

Global Context Enhanced Social Recommendation with Hierarchical Graph Neural Networks

A Graph Neural Network Framework for Social Recommendations

Exploiting Multifaceted Nature of Items and Users for Session-based Recommendation

Multi-Aspect enhanced Graph Neural Networks for recommendation

Enhancing Hierarchy-Aware Graph Networks with Deep Dual Clustering for Session-based Recommendation

Hierarchical Hyperedge Embedding-Based Representation Learning for Group Recommendation

Reviews Meet Graphs: Enhancing User and Item Representations for Recommendation with Hierarchical Attentive Graph Neural Network.

Enhancing Recommendation with Automated Tag Taxonomy Construction in Hyperbolic Space

Disentangled Hypergraph Collaborative Filtering for Social Recommendation

HRCF: Enhancing Collaborative Filtering Via Hyperbolic Geometric Regularization