arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2203.09553cs.AIcs.CLcs.LG
收藏

基于隐私保护关系嵌入聚合的知识图谱高效联邦学习

Efficient Federated Learning on Knowledge Graphs via Privacy-preserving Relation Embedding Aggregation

Lehigh University(理海大学);University of Illinois Chicago(伊利诺伊大学芝加哥分校);Carnegie Mellon University(卡内基梅隆大学);Virginia Tech(弗吉尼亚理工大学);Emory University(埃默里大学);Samsung Research America(三星美国研究院)

Kai Zhang, Yu Wang, Hongyi Wang, Lifu Huang, Carl Yang, Xun Chen, Lichao Sun

摘要:

联邦学习(FL)在多源知识图谱(KG)上的知识表示、推理和数据采掘应用中是必不可少的。最近的一项研究FedE首次提出了一个FL框架,该框架在所有客户端之间共享KG的实体嵌入。然而,从FedE实体嵌入共享将招致严重的隐私泄漏。具体地,可以使用已知实体嵌入来推断私有客户端中是否存在两个实体之间的特定关系。在本文中,我们介绍了一种新的攻击方法,旨在恢复的基础上嵌入信息的原始数据,这是进一步用于评估FedE的脆弱性。此外,我们提出了一个联邦学习范式与隐私保护的关系嵌入聚合(FedR),以解决隐私问题的FedE。此外,关系嵌入共享由于其较小的查询大小,可以显着降低通信成本。我们进行了广泛的实验,以评估FedR与五个不同的KG嵌入模型和三个数据集。与FedE相比,FedR在链接预测任务上实现了相似的效用,并在隐私保护效果和通信效率方面有了显著的提高。

英文摘要:

Federated learning (FL) can be essential in knowledge representation, reasoning, and data mining applications over multi-source knowledge graphs (KGs). A recent study FedE first proposes an FL framework that shares entity embeddings of KGs across all clients. However, entity embedding sharing from FedE would incur a severe privacy leakage. Specifically, the known entity embedding can be used to infer whether a specific relation between two entities exists in a private client. In this paper, we introduce a novel attack method that aims to recover the original data based on the embedding information, which is further used to evaluate the vulnerabilities of FedE. Furthermore, we propose a Federated learning paradigm with privacy-preserving Relation embedding aggregation (FedR) to tackle the privacy issue in FedE. Besides, relation embedding sharing can significantly reduce the communication cost due to its smaller size of queries. We conduct extensive experiments to evaluate FedR with five different KG embedding models and three datasets. Compared to FedE, FedR achieves similar utility and significant improvements regarding privacy-preserving effect and communication efficiency on the link prediction task.

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。

↑