Abstract:Cross-domain recommendation (CDR) aims to improve recommendation accuracy in sparse domains by transferring knowledge from data-rich domains. However, existing CDR methods often assume the availability of user-item interaction data across domains, overlooking user privacy concerns. Furthermore, these methods suffer from performance degradation in scenarios with sparse overlapping users, as they typically depend on a large number of fully shared users for effective knowledge transfer. To address these challenges, we propose a Federated Prototype-based Contrastive Learning (CL) method for Privacy-Preserving CDR, named FedPCL-CDR. This approach utilizes non-overlapping user information and prototypes to improve multi-domain performance while protecting user privacy. FedPCL-CDR comprises two modules: local domain (client) learning and global server aggregation. In the local domain, FedPCL-CDR clusters all user data to learn representative prototypes, effectively utilizing non-overlapping user information and addressing the sparse overlapping user issue. It then facilitates knowledge transfer by employing both local and global prototypes returned from the server in a CL manner. Simultaneously, the global server aggregates representative prototypes from local domains to learn both local and global prototypes. The combination of prototypes and federated learning (FL) ensures that sensitive user data remains decentralized, with only prototypes being shared across domains, thereby protecting user privacy. Extensive experiments on four CDR tasks using two real-world datasets demonstrate that FedPCL-CDR outperforms the state-of-the-art baselines.

What problem does this paper attempt to address?

### What problem does this paper attempt to solve? This paper aims to address two main challenges in Cross-domain Recommendation (CDR): 1. **How to effectively protect user privacy during cross-domain knowledge transfer?** - Existing CDR methods typically assume that user-item interaction data can be directly transferred between different domains, which poses serious privacy issues. Most existing methods rely on fully overlapping users to achieve knowledge transfer, but these methods often fail to adequately protect user privacy in practical applications. 2. **How to improve recommendation performance when the cross-domain user overlap is low?** - Many existing CDR methods depend on fully overlapping users as a bridge to transfer knowledge. However, in real-world datasets, the number of overlapping users is very small. For example, the proportion of overlapping users in the Amazon dataset is only 5%. In such cases, model performance significantly declines. Therefore, it is necessary to utilize information from non-overlapping users to improve recommendation performance. To address these challenges, the authors propose a method based on federated learning and prototype contrastive learning, called FedPCL-CDR. This method addresses the issues in the following ways: - **Utilizing non-overlapping user information**: By clustering all user data to generate representative prototypes, it effectively utilizes information from non-overlapping users, addressing the issue of sparse overlapping users. - **Protecting user privacy**: Within the federated learning framework, only prototypes are shared instead of raw user data, ensuring that sensitive user data remains decentralized, thereby protecting user privacy. - **Dual prototype contrastive learning**: By conducting contrastive learning with both local and global prototypes, it achieves effective cross-domain knowledge transfer, enhancing the performance of the recommendation system. Through these methods, FedPCL-CDR demonstrates outstanding performance in multiple CDR tasks, surpassing existing baseline methods.

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

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

Privacy-Preserving Cross-Domain Recommendation with Federated Graph Learning

Win-Win: A Privacy-Preserving Federated Framework for Dual-Target Cross-Domain Recommendation.

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

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

Federated Graph Learning for Cross-Domain Recommendation

Differential Private Knowledge Transfer for Privacy-Preserving Cross-Domain Recommendation

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

Federated Probabilistic Preference Distribution Modelling with Compactness Co-Clustering for Privacy-Preserving Multi-Domain Recommendation

FedCL: Federated Contrastive Learning for Privacy-Preserving Recommendation

A Personalized Cross-Domain Recommendation with Federated Meta Learning

Contrastive Cross-domain Recommendation in Matching

FedHCDR: Federated Cross-Domain Recommendation with Hypergraph Signal Decoupling

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

PPGenCDR: A Stable and Robust Framework for Privacy-Preserving Cross-Domain Recommendation

Reducing Item Discrepancy Via Differentially Private Robust Embedding Alignment for Privacy-Preserving Cross Domain Recommendation

Differentially Private Sparse Mapping for Privacy-Preserving Cross Domain Recommendation

FedPDD: A Privacy-preserving Double Distillation Framework for Cross-silo Federated Recommendation

An Enhanced Federated Prototype Learning Method under Domain Shift

Transferring Causal Mechanism over Meta-representations for Target-unknown Cross-domain Recommendation