Multi-view semantic enhancement model for few-shot knowledge graph completion

Ruixin Ma, Hao Wu, Xiaoru Wang, Weihe Wang, Yunlong Ma, Liang Zhao

Abstract: Highlights•We design a multi-view joint encoder to learn semantic information.•Commonsense semantic constraints are introduced to explore semantic relations.•We leverage the commonsense view to generate high-quality negative triples.•We identify candidate entities that match commonsense in the prediction process.