EMNLPNov, 2020

DyERNIE: Riemann 流形嵌入的动态演化,用于时间知识图谱补全

TL;DRDy-ERNIE proposes a non-Euclidean embedding approach that learns evolving entity representations on a product of Riemannian manifolds to better reflect the geometric structures of temporal knowledge graphs, resulting in improved performance on temporal knowledge graph completion tasks.