arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2310.20193cs.LGcs.CRcs.IR
收藏

FedRec+:增强联邦推荐系统中的隐私并应对异质性

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

The Chinese University of Hong Kong (Shenzhen)(香港中文大学(深圳))

Lin Wang, Zhichao Wang, Xi Leng, Xiaoying Tang

AI总结:

本文提出联邦推荐集成框架FedRec+,通过基于特征相似性的最优子集选择生成虚拟评分以增强隐私并降低噪声,同时利用Wasserstein距离估计客户端异质性和贡献以推导最优聚合权重,在多个数据集上取得最先进性能。

摘要:

在推荐系统中,保护边缘用户隐私并降低通信成本面临重大挑战。尽管联邦学习通过避免客户端与服务器之间的数据交换,已被证明在保护隐私方面有效,但已有研究表明,服务器可以基于用户连续两轮上传梯度中获得的非零更新梯度推断用户评分。此外,联邦推荐系统(FRS)还面临异质性挑战,导致推荐性能下降。本文提出FedRec+,一个用于FRS的集成框架,在增强隐私的同时应对异质性挑战。FedRec+基于特征相似性进行最优子集选择,仅利用用户本地信息为伪物品生成近最优虚拟评分。该方法在不增加额外通信成本的情况下降低噪声。此外,我们利用Wasserstein距离估计每个客户端的异质性和贡献,并通过求解一个定义的优化问题推导最优聚合权重。实验结果表明,FedRec+在多个参考数据集上取得了最先进的性能。

英文摘要:

Preserving privacy and reducing communication costs for edge users pose significant challenges in recommendation systems. Although federated learning has proven effective in protecting privacy by avoiding data exchange between clients and servers, it has been shown that the server can infer user ratings based on updated non-zero gradients obtained from two consecutive rounds of user-uploaded gradients. Moreover, federated recommendation systems (FRS) face the challenge of heterogeneity, leading to decreased recommendation performance. In this paper, we propose FedRec+, an ensemble framework for FRS that enhances privacy while addressing the heterogeneity challenge. FedRec+ employs optimal subset selection based on feature similarity to generate near-optimal virtual ratings for pseudo items, utilizing only the user's local information. This approach reduces noise without incurring additional communication costs. Furthermore, we utilize the Wasserstein distance to estimate the heterogeneity and contribution of each client, and derive optimal aggregation weights by solving a defined optimization problem. Experimental results demonstrate the state-of-the-art performance of FedRec+ across various reference datasets.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑