Federated Knowledge Graph

Federated Knowledge Graph (FKGE) research focuses on collaboratively training knowledge graph embedding models across multiple, distributed datasets without directly sharing sensitive data. Current efforts concentrate on improving communication efficiency through techniques like embedding sparsification and knowledge distillation, as well as addressing data heterogeneity and ensuring model personalization across diverse clients. This field is significant because it enables large-scale knowledge graph construction and reasoning while preserving data privacy, with applications ranging from bioinformatics to more general large language model data retrieval.