Personalized Federated Knowledge Graph Embedding with Client-Wise Relation Graph

Xiaoxiong Zhang,Zhiwei Zeng,Xin Zhou,Dusit Niyato,Zhiqi Shen

2024-06-18

Abstract:Federated Knowledge Graph Embedding (FKGE) has recently garnered considerable interest due to its capacity to extract expressive representations from distributed knowledge graphs, while concurrently safeguarding the privacy of individual clients. Existing FKGE methods typically harness the arithmetic mean of entity embeddings from all clients as the global supplementary knowledge, and learn a replica of global consensus entities embeddings for each client. However, these methods usually neglect the inherent semantic disparities among distinct clients. This oversight not only results in the globally shared complementary knowledge being inundated with too much noise when tailored to a specific client, but also instigates a discrepancy between local and global optimization objectives. Consequently, the quality of the learned embeddings is compromised. To address this, we propose Personalized Federated knowledge graph Embedding with client-wise relation Graph (PFedEG), a novel approach that employs a client-wise relation graph to learn personalized embeddings by discerning the semantic relevance of embeddings from other clients. Specifically, PFedEG learns personalized supplementary knowledge for each client by amalgamating entity embedding from its neighboring clients based on their "affinity" on the client-wise relation graph. Each client then conducts personalized embedding learning based on its local triples and personalized supplementary knowledge. We conduct extensive experiments on four benchmark datasets to evaluate our method against state-of-the-art models and results demonstrate the superiority of our method.

Information Retrieval,Artificial Intelligence

What problem does this paper attempt to address?

This paper attempts to address the issue of semantic differences in Federated Knowledge Graph Embedding (FKGE). Specifically, existing FKGE methods often overlook the inherent semantic differences between Knowledge Graphs (KGs) of different clients, resulting in global supplementary knowledge containing excessive noise for specific clients and causing inconsistencies between local and global optimization objectives, thereby reducing the quality of the learned embeddings.

To solve this problem, the authors propose Personalized Federated Knowledge Graph Embedding with Client-Wise Relation Graph (PFedEG). PFedEG leverages a client-wise relation graph to learn personalized embeddings for each client. The specific method involves integrating entity embeddings based on the client-wise relation graph to generate personalized supplementary knowledge for each client. This approach allows each client to perform personalized embedding learning based on its local triples and personalized supplementary knowledge, thereby improving embedding quality. Experimental results on 4 benchmark datasets demonstrate that this method outperforms existing state-of-the-art models.