Abstract:Federated recommendation systems employ federated learning techniques to safeguard user privacy by transmitting model parameters instead of raw user data between user devices and the central server. Nevertheless, the current federated recommender system faces three significant challenges: (1) data heterogeneity: the heterogeneity of users’ attributes and local data necessitates the acquisition of personalized models to improve the performance of federated recommendation; (2) model performance degradation: the privacy-preserving protocol design in the federated recommendation, such as pseudo item labeling and differential privacy, would deteriorate the model performance; (3) communication bottleneck: the standard federated recommendation algorithm can have a high communication overhead. Previous studies have attempted to address these issues, but none have been able to solve them simultaneously. In this paper, we propose a novel framework, named PerFedRec++ , to enhance the personalized federated recommendation with self-supervised pre-training. Specifically, we utilize the privacy-preserving mechanism of federated recommender systems to generate two augmented graph views, which are used as contrastive tasks in self-supervised graph learning to pre-train the model. Pre-training enhances the performance of federated models by improving the uniformity of representation learning. Also, by providing a better initial state for federated training, pre-training makes the overall training converge faster, thus alleviating the heavy communication burden. We then construct a collaborative graph to learn the client representation through a federated graph neural network. Based on these learned representations, we cluster users into different user groups and learn personalized models for each cluster. Each user learns a personalized model by combining the global federated model, the cluster-level federated model, and its own fine-tuned local model. Experiments on three real-world datasets show that our proposed method achieves superior performance over existing methods.

AeroRec: an Efficient On-Device Recommendation Framework Using Federated Self-Supervised Knowledge Distillation

ReFRS: Resource-efficient Federated Recommender System for Dynamic and Diversified User Preferences

Efficient and Robust Regularized Federated Recommendation

FedRec++: Lossless Federated Recommendation with Explicit Feedback

Efficient Federated Recommender System with Adaptive Model Pruning and Momentum-based Batch Adjustment

Fast-adapting and privacy-preserving federated recommender system

FedRec+: Enhancing Privacy and Addressing Heterogeneity in Federated Recommendation Systems

FedRec: Federated Recommendation With Explicit Feedback

FedRec: Privacy-Preserving News Recommendation with Federated Learning

Federated Low-Rank Tensor Projections for Sequential Recommendation

Towards Communication Efficient and Fair Federated Personalized Sequential Recommendation

Federated Unlearning for On-Device Recommendation

Efficient-FedRec: Efficient Federated Learning Framework for Privacy-Preserving News Recommendation

A split-federated learning and edge-cloud based efficient and privacy-preserving large-scale item recommendation model

FedDeepFM: A Factorization Machine-Based Neural Network for Recommendation in Federated Learning

Enhancing Recommendation Capabilities Using Multi-Head Attention-Based Federated Knowledge Distillation

Towards fair and personalized federated recommendation

LightFR: Lightweight Federated Recommendation with Privacy-preserving Matrix Factorization

PDC-FRS: Privacy-preserving Data Contribution for Federated Recommender System

PerFedRec++: Enhancing Personalized Federated Recommendation with Self-Supervised Pre-Training

HeteFedRec: Federated Recommender Systems with Model Heterogeneity