.png){kind=icon}
.png){kind=icon}
Return
FuST-KGC: Fusing sub-graph structures and textual semantics for knowledge graph completion
.png){kind=icon}
PRE
AI
DOI：10.1016/j.neucom.2026.132690.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
• We propose FuST-KGC, a novel method that fuses subgraph structures and textual semantics for knowledge graph completion. • We introduce a dynamic, degree-guided sampling strategy to construct informative local subgraphs that effectively capture crucial reasoning paths. • FuST-KGC linearizes multi-hop subgraph topology into a text sequence, enabling a single pre-trained language model to fuse structural context with fine-grained semantics. • FuST-KGC addresses the limitations of existing models that oversimplify graph structures or use complex, separate encoders, improving both reasoning accuracy and semantic-structural integration.