Uncertain Knowledge Graphs (UKGs) have recently garnered significant attention, primarily for their capabilities in modeling the inherent uncertainty of relation facts in a knowledge graph with a confidence score and reasoning based on uncertainty knowledge embeddings. Existing uncertain knowledge graph embedding (UKGE) methods ignore the uncertainty of neighbor entities in the modeling of graph structure-aware context, on the other hand, they have inadequately accounted for the confidence information of relational triples when learning their embeddings. To address these issues, in this paper, we propose a novel uncertainty knowledge graph embedding model, named UnKGE-GSFC, which works to learn graph structure-aware context and fusedtriple confidence information in an encoder-decoder framework. To be specific, for encoding, we design an uncertainty-aware RGCN providing foundations for learning graph structure-aware context. For decoding, we model confidence for each triple by fusing its semantic information and effectively preserving sequential dependencies. Extensive experiments over publicly available UKGE benchmarks demonstrate that our approach can significantly outperform previous state-of-the-art methods in different evaluation metrics on confidence prediction and link prediction tasks, respectively.
Song et al. (Thu,) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: