Abstract:In recent years, Cross-Domain Recommendation (CDR) has drawn significant attention, which utilizes user data from multiple domains to enhance the recommendation performance. However, current CDR methods require sharing user data across domains, thereby violating the General Data Protection Regulation (GDPR). Consequently, numerous approaches have been proposed for Federated Cross-Domain Recommendation (FedCDR). Nevertheless, the data heterogeneity across different domains inevitably influences the overall performance of federated learning. In this study, we propose FedHCDR, a novel Federated Cross-Domain Recommendation framework with Hypergraph signal decoupling. Specifically, to address the data heterogeneity across domains, we introduce an approach called hypergraph signal decoupling (HSD) to decouple the user features into domain-exclusive and domain-shared features. The approach employs high-pass and low-pass hypergraph filters to decouple domain-exclusive and domain-shared user representations, which are trained by the local-global bi-directional transfer algorithm. In addition, a hypergraph contrastive learning (HCL) module is devised to enhance the learning of domain-shared user relationship information by perturbing the user hypergraph. Extensive experiments conducted on three real-world scenarios demonstrate that FedHCDR outperforms existing baselines significantly.

What problem does this paper attempt to address?

### Problems the Paper Attempts to Solve The paper attempts to address the issue of how to effectively handle data heterogeneity between different domains in Cross-Domain Recommendation (CDR) while protecting user privacy. Specifically, the paper focuses on the problem of Federated Cross-Domain Recommendation (FedCDR). ### Background and Challenges 1. **Cross-Domain Recommendation (CDR)**: - Utilizes user data from multiple domains to improve the performance of recommendation systems. - Assumes that user preferences are similar across different domains, enhancing recommendation effects by sharing user rating data. 2. **Privacy Protection Issue**: - With the implementation of the General Data Protection Regulation (GDPR), user item rating data cannot be directly shared between different domains. - Providing high-quality cross-domain recommendations while protecting privacy has become an urgent problem to solve. 3. **Federated Cross-Domain Recommendation (FedCDR)**: - In this context, user item rating interactions are considered private information and cannot be directly accessed by other domains. - Existing FedCDR methods, while addressing privacy issues, face the problem of data heterogeneity, where user item interaction data in different domains contains domain-specific information. ### Data Heterogeneity Problem - **Data Heterogeneity**: User item interaction data in different domains contains domain-specific information, which affects the overall performance of federated learning. - **Limitations of Existing Methods**: Existing FedCDR methods transmit knowledge by directly aggregating client models or user representations, leading to domain-specific information being mixed into the global model, thus affecting local performance (i.e., negative transfer). ### Solution To address the data heterogeneity problem, the paper proposes a new federated cross-domain recommendation framework—FedHCDR (Federated Cross-Domain Recommendation with Hypergraph Signal Decoupling). The main contributions of this framework include: 1. **Hypergraph Signal Decoupling (HSD)**: - Uses high-pass and low-pass hypergraph filters to decouple user features into domain-specific and domain-shared features. - Trains these filters through a local-global bidirectional transmission algorithm to extract domain-specific and domain-shared user representations. 2. **Hypergraph Contrastive Learning Module (HCL)**: - Enhances the learning of domain-shared user relationship information by perturbing the user hypergraph. ### Experimental Validation - **Experimental Setup**: Extensive experiments were conducted on the Amazon dataset to validate the performance of FedHCDR in three different FedCDR scenarios. - **Experimental Results**: The results show that FedHCDR significantly outperforms existing baseline methods, especially in metrics such as MRR, NDCG@10, and HR@10. ### Conclusion The FedHCDR framework proposed in the paper effectively addresses the data heterogeneity problem in federated cross-domain recommendation through hypergraph signal decoupling and hypergraph contrastive learning, while ensuring the protection of user data privacy. Experimental results validate the effectiveness and superiority of this method.

FedHCDR: Federated Cross-Domain Recommendation with Hypergraph Signal Decoupling

Federated Graph Learning for Cross-Domain Recommendation

User Distribution Mapping Modelling with Collaborative Filtering for Cross Domain Recommendation

Heterogeneous Graph-based Framework with Disentangled Representations Learning for Multi-target Cross Domain Recommendation

Privacy-Preserving Cross-Domain Recommendation with Federated Graph Learning

Prompt-enhanced Federated Content Representation Learning for Cross-domain Recommendation

Federated Prototype-based Contrastive Learning for Privacy-Preserving Cross-domain Recommendation

Decoupled domain-specific and domain-conditional representation learning for cross-domain recommendation

Heterogeneous and Clustering-Enhanced Personalized Preference Transfer for Cross-Domain Recommendation

Enhancing Dual-Target Cross-Domain Recommendation with Federated Privacy-Preserving Learning

Intra and Inter Domain HyperGraph Convolutional Network for Cross-Domain Recommendation

DTCDR: A Framework for Dual-Target Cross-Domain Recommendation

Contrastive Cross-domain Recommendation in Matching

Beyond the Overlapping Users: Cross-Domain Recommendation via Adaptive Anchor Link Learning

A Privacy-Preserving Framework with Multi-Modal Data for Cross-Domain Recommendation

A Centralized-Distributed Transfer Model for Cross-Domain Recommendation Based on Multi-Source Heterogeneous Transfer Learning

FedRL: A Reinforcement Learning Federated Recommender System for Efficient Communication Using Reinforcement Selector and Hypernet Generator

Towards Universal Cross-Domain Recommendation

Hyperbolic Knowledge Transfer in Cross-Domain Recommendation System

DDGHM: Dual Dynamic Graph with Hybrid Metric Training for Cross-Domain Sequential Recommendation

Federated User Preference Modeling for Privacy-Preserving Cross-Domain Recommendation