![[2602.21957] Learning to Collaborate via Structures: Cluster-Guided Item Alignment for Federated Recommendation](https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png){kind=logo}
[2602.21957] Learning to Collaborate via Structures: Cluster-Guided Item Alignment for Federated Recommendation
Summary
The paper presents CGFedRec, a novel framework for federated recommendation that enhances collaboration by using cluster-guided item alignment, improving communication efficiency and recommendation accuracy.
Why It Matters
As federated learning becomes increasingly important for privacy-preserving machine learning, this research offers a significant advancement in recommendation systems by reducing the need for high-dimensional data transmission while still capturing user preferences effectively. This can lead to better user experiences in applications like e-commerce and content recommendation.
Key Takeaways
- CGFedRec improves federated recommendation by using cluster labels instead of full item embeddings.
- The framework enhances communication efficiency while maintaining high recommendation accuracy.
- It allows for local variation in item representations, promoting user personalization.
- Global semantic relationships among items serve as structural constraints for better collaboration.
- Extensive experiments validate the effectiveness of the proposed approach across various datasets.
Computer Science > Information Retrieval arXiv:2602.21957 (cs) [Submitted on 25 Feb 2026] Title:Learning to Collaborate via Structures: Cluster-Guided Item Alignment for Federated Recommendation Authors:Yuchun Tu, Zhiwei Li, Bingli Sun, Yixuan Li, Xiao Song View a PDF of the paper titled Learning to Collaborate via Structures: Cluster-Guided Item Alignment for Federated Recommendation, by Yuchun Tu and 4 other authors View PDF HTML (experimental) Abstract:Federated recommendation facilitates collaborative model training across distributed clients while keeping sensitive user interaction data local. Conventional approaches typically rely on synchronizing high-dimensional item representations between the server and clients. This paradigm implicitly assumes that precise geometric alignment of embedding coordinates is necessary for collaboration across clients. We posit that establishing relative semantic relationships among items is more effective than enforcing shared representations. Specifically, global semantic relations serve as structural constraints for items. Within these constraints, the framework allows item representations to vary locally on each client, which flexibility enables the model to capture fine-grained user personalization while maintaining global consistency. To this end, we propose Cluster-Guided FedRec framework (CGFedRec), a framework that transforms uploaded embeddings into compact cluster labels. In this framework, the server functions as a global ...
