Cluster-Enhanced Federated Graph Neural Network for Recommendation

TL;DR

This paper introduces CFedGR, a federated GNN framework that enhances user graphs with high-order signals via clustering, improving recommendation accuracy while preserving user privacy.

Contribution

The paper proposes a novel federated GNN method that incorporates high-order collaborative signals through clustering, reducing communication costs and enhancing privacy in recommender systems.

Findings

Effective augmentation of user graphs with high-order signals.

Reduced communication overhead in federated training.

Improved recommendation performance on benchmark datasets.

Abstract

Personal interaction data can be effectively modeled as individual graphs for each user in recommender systems.Graph Neural Networks (GNNs)-based recommendation techniques have become extremely popular since they can capture high-order collaborative signals between users and items by aggregating the individual graph into a global interactive graph.However, this centralized approach inherently poses a threat to user privacy and security. Recently, federated GNN-based recommendation techniques have emerged as a promising solution to mitigate privacy concerns. Nevertheless, current implementations either limit on-device training to an unaccompanied individual graphs or necessitate reliance on an extra third-party server to touch other individual graphs, which also increases the risk of privacy leakage. To address this challenge, we propose a Cluster-enhanced Federated Graph Neural Network framework for Recommendation, named CFedGR, which introduces high-order collaborative signals to augment individual graphs in a privacy preserving manner. Specifically, the server clusters the pretrained user representations to identify high-order collaborative signals. In addition, two efficient strategies are devised to reduce communication between devices and the server. Extensive experiments on three benchmark datasets validate the effectiveness of our proposed methods.